U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Committee on Prepositioned Medical Countermeasures for the Public; Institute of Medicine; Stroud C, Viswanathan K, Powell T, et al., editors. Prepositioning Antibiotics for Anthrax. Washington (DC): National Academies Press (US); 2011 Sep 30.

Cover of Prepositioning Antibiotics for Anthrax

Prepositioning Antibiotics for Anthrax.

Show details

4Prepositioning Strategies

Policy makers are considering prepositioning to complement existing more centralized strategies because prepositioning strategies may serve to:

  • increase the number of potentially exposed people who are able to access antibiotics within an appropriate time frame following an anthrax attack;
  • decrease the burden on existing strategies for dispensing medical countermeasures (MCM), especially the use of points of dispensing (PODs), and reduce surge demand on the health care system; and
  • enhance fairness and equitability in access to antibiotics.

As discussed in this chapter and further in Chapter 5, however, these strategies also can be associated with higher levels of inappropriate use and health risks, higher costs, and additional practical burdens relative to existing strategies.

Antibiotics may be prepositioned in many different venues using many different strategies, including:

  • forward-deployed MCM—MCM stored near the locations from which they will be dispensed,
  • cached MCM—MCM stored at the locations from which they will be dispensed,1 and
  • predispensed MCM—MCM stored by the intended users or by heads of households or other nonmedical caregivers for use by those in their care.

This chapter describes these three categories of prepositioning strategies. For each category, several example strategies, the potential roles for those strategies within a jurisdiction's overall dispensing strategy, and potential health risks, if any, are discussed. Also discussed for each category are practical considerations, including logistics, communication needs, expected behavior and adherence, and legal and regulatory issues. Table 4-1 summarizes key features of the three categories of prepositioning strategies. This table is not intended to be comprehensive; other push and pull strategies (such as those described in Chapter 3) also could be employed to enhance distribution and dispensing.

TABLE 4-1. Practical Considerations for Developing Prepositioning Strategies.


Practical Considerations for Developing Prepositioning Strategies.

In this chapter, the committee discusses the individual properties of different prepositioning strategies to highlight the specific uses of each and the associated advantages, disadvantages, and other considerations. However, these strategies are likely to be used in combination not only during initial prophylaxis, but also later when it is necessary to provide the exposed population with vaccine and a prolonged antibiotic course. This chapter focuses primarily on the qualitative considerations that should factor into jurisdictions' decisions about whether to develop strategies for prepositioning prophylactic antibiotics in their communities. Chapter 5 outlines a decision-aiding framework and a model for quantifying and comparing health benefits and economic costs across the various prepositioning strategies and presents the committee's recommendations on this topic.


Forward-deployed MCM are stored near the locations where they will be dispensed. The primary purpose of forward-deploying MCM is to decrease the transportation time associated with distributing the MCM from stockpiles to PODs. Several entities could potentially maintain forward-deployed stockpiles of antibiotics, including the Centers for Disease Control and Prevention (CDC)/Strategic National Stockpile (SNS); other federal agencies, such as the Department of Veteran Affairs (VA) and Department of Defense (DOD); state and local authorities; and commercial pharmaceutical distributors. These strategies are described below.

Forward-Deployed by the Strategic National Stockpile

SNS stockpiles currently are held in large, strategically placed warehouses throughout the nation. For security reasons, the locations of SNS stockpiles are not disclosed. Therefore, it is possible—and perhaps likely—that certain SNS warehouses are located near high-risk areas, such as major cities, and therefore would already be considered forward-deployed. The SNS could be further forward-deployed by establishing additional SNS warehouses. This might be done, for example, by prepositioning MCM in SNS-managed warehouses in the 11 Tier 1 cities of the Urban Areas Security Initiative, which are the metropolitan areas that the Department of Homeland Security (DHS) has determined to be at highest risk of a terrorist attack (Burel, 2011; DHS, 2011).

The primary motivation for forward-deploying SNS stockpiles is to decrease the time associated with transportation from the SNS warehouse to state receiving, staging, and storing (RSS) warehouses, which then redistribute the MCM to the jurisdictions' PODs. This strategy would enable PODs to begin dispensing antibiotics more quickly, thereby increasing the number of people receiving prophylactic antibiotics within the time window in which they can prevent anthrax. Decreasing the transportation time from SNS warehouses to state RSS sites will be effective, however, only if the RSS sites and PODs can be set up and staffed quickly enough to take advantage of the reduced delivery time (Burel, 2011). If MCM are delivered from the SNS before the RSS sites are ready to redistribute them or are redistributed from RSS sites to PODs before the PODs are ready to begin dispensing, the reduced delivery time from SNS warehouses will have no impact on the time at which dispensing of the MCM begins. Although data are sparse on the time currently required for state and local authorities to commence POD operations, and this time is likely to show great variability across jurisdictions, the limited data available suggest that 8 hours or more may be needed (Burel, 2011). Therefore, decreasing the SNS transportation time to under 8 hours is unlikely to be cost-effective unless jurisdictions can set up PODs more quickly. For those states and localities that already have the ability to set up RSS sites and PODs rapidly, reducing the SNS delivery time to 8 hours or less might induce some state and local entities to consider eliminating or reducing the quantity of antibiotics in their caches as a cost-saving measure.

Forward-deploying MCM that remain under the control of the SNS (rather than transferring them to state, local, or private entities) would decrease transportation time while still enabling central coordination by the SNS; some flexibility to use the SNS infrastructure to redeploy MCM to other areas of need; and the use of the Shelf Life Extension Program (SLEP, described in Chapter 3), which is available only to selected federal stockpiles (Courtney et al., 2009).

Forward-deploying SNS stockpiles would require storing MCM in more locations compared with storage in fewer, more centralized warehouses; therefore, forward-deployment would impose a higher management burden and require a greater quantity of medication, with associated costs. This strategy also would decrease flexibility to reallocate antibiotics if an attack occurred in a location with lower perceived risk.

CDC's CHEMPACK project is an example of forward-deployed SNS materiel (Box 4-1). The most significant difference between an attack with a nerve agent and anthrax is the time frame postexposure within which MCM are effective: antidotes for nerve agents must be administered within minutes to hours, compared with several days for anthrax antibiotics. CDC also forward-deploys DTPA (diethylene triamine pentaacetic acid) in states to reduce response time in the event of a radiological incident. In 2010, CDC reported that “as of March 2010, 89% of the 62 [Department of Health and Human Services–funded] state, locality, and U.S. insular area public health departments received 78,880 doses of [DTPA] from CDC's [SNS]” (CDC, 2010d, p. 28).

Box Icon

BOX 4-1

CHEMPACK: Forward-Deployed Strategic National Stockpile (SNS) Antidotes for Nerve Agents. Nerve agents (e.g., ricin, sarin gas) can be absorbed through the eyes or skin, ingested by eating or drinking contaminated food or water, and inhaled; they can (more...)

Forward-Deployed by Other Federal Agencies

MCM also could be forward-deployed at VA hospitals or DOD medical treatment facilities, which are located throughout the country. Although these activities are beyond the normal scope of the VA and DOD mandates, the potential to use this health system infrastructure should not be overlooked. These facilities already maintain pharmaceutical caches for their staff and patients and have the medical staff and infrastructure required to monitor and store medications properly (VA, 2010; see Appendix D). The SNS already partners with the VA and DOD, and the VA currently provides the SNS with acquisition support (CDC, 2010a; VA, 2011). Therefore, instead of using resources to establish new SNS warehouses, it might be possible to expand VA or DOD caches to include antibiotics for the public. These MCM would be distributed to open PODs for dispensing; VA or DOD facilities would be unlikely to serve as open PODs since they would be occupied with providing health care following an attack. Stockpiles at VA or DOD facilities would allow use of the SLEP to minimize costs associated with expiration, and potentially could even be cycled through regular health care uses to avoid expiration entirely.

Forward-Deployed by State Authorities

MCM could be forward-deployed by states to locations at high risk instead of being kept in a single central location within the state. For example, Minnesota maintains 11 caches of MCM throughout the state, distributed according to population density and proximity to major cities (Minnesota Department of Health, 2006).

In 2005, the New York State Department of Health's Office of Health Emergency Preparedness established all-hazard medical emergency response caches (MERCs) in multiple locations throughout the state (NYSOHS, 2007). The MERCs contain pharmaceuticals and devices (e.g., doxycycline and ciprofloxacin, Mark I Autoinjector Kits with antidotes to nerve agents) and other medical supplies and equipment that can be deployed rapidly in an emergency, reaching any area in the state within 4 hours (compared with 12 hours in the case of SNS Push Packages). The state-owned assets in the MERCs are the first supplies available to move into an affected area. Depending on the incident, if these state supplies are committed or expended, the state requests assistance from the SNS. The MERCs also eliminate the need for multiple local stockpiles that are not cost-effective. The MERCs are intended to bridge the supply gap between the time of an incident and the arrival and distribution of SNS resources.

Like the SNS forward-deployed strategy, this strategy offers the advantage of decreased transportation time to POD sites. It also has similar drawbacks, including increased costs associated with supporting more facilities and potentially greater quantities of MCM, and decreased flexibility to reassign MCM should an attack occur at a location other than those predicted. Unlike forward-deploying at SNS warehouses or in VA or DOD facilities, however, state stockpiles cannot take advantage of the SLEP (Courtney et al., 2009). Although this is a limitation for centralized state stockpiles as well, it could have a greater impact on disbursed forward-deployed stockpiles since they would likely require a greater quantity of antibiotics overall, and the administrative burden of monitoring, discarding, replacing, or cycling medications through regular health care uses would increase as the number of stockpile locations increased. In the past, states attempting to contract with pharmaceutical distributors to rotate the antibiotics in their stockpiles have faced the challenge of a low market demand for the MCM, and thus little benefit in terms of decreasing replacement costs due to expiring medications (Courtney et al., 2009). This might be an area for which national guidance would be beneficial, as discussed in detail later in this chapter.

Several studies have suggested that state and local planning efforts should focus first on increasing dispensing capacity because increasing local inventory is cost-effective and effective in reducing mortality if the community already has a highly robust dispensing capacity (Bravata et al., 2006; Zaric et al., 2008). This issue is examined further in Chapter 5.

Forward-Deployed by Commercial Entities

MCM could be forward-deployed by commercial pharmaceutical distributors, including both companies that specialize primarily in supply chain management and pharmaceutical distribution (e.g., McKesson, Cardinal Health, AmerisourceBergen) and companies that distribute pharmaceuticals to supply their retail stores (e.g., Target, Walmart). Commercial pharmaceutical distributors could forward-deploy MCM on behalf of either public health authorities to supply open PODs or private-sector entities to supply closed PODs for employees and their families. Under this strategy, a jurisdiction or private-sector entity would contract with a pharmaceutical distributor to maintain stockpiles of antibiotics in a warehouse near the anticipated POD location(s). This strategy would limit the time needed for transportation (relative to relying on pharmaceutical warehouses located far away) and would take advantage of these distributors' expertise in supply chain management and medication storage. The challenge for public health authorities would be to identify pharmaceutical distributors that were interested in participating in such a program and would find it worthwhile from a business perspective. While distributors may be willing to store MCM in bulk, they are likely to find it more challenging to store MCM in prelabeled unit-of-use quantities because of storage space requirements and costs. CDC currently is exploring the use of pharmaceutical distributors to distribute antiviral drugs from the SNS (CDC, 2011a).

During the 2009 H1N1 influenza pandemic, the state of Virginia employed this model, contracting with a private vendor (AmerisourceBergen) to store and distribute its supply of antivirals at the direction of the state health commissioner (Virginia Department of Health, 2009). Virginia is pursuing additional partnerships with chains and pharmacies based on the system set up in response to the 2009 pandemic.


Cached MCM are positioned in the locations from which they will be dispensed. The caches may be located in health care facilities (e.g., hospitals and pharmacies) or non–health care facilities (e.g., non–health care workplaces) and may be maintained by public or private entities. The specific purposes, advantages, and challenges involved depend on the type of cache and are described below.

Caches in Health Care Settings

The primary purpose of MCM caches in health care settings is to distribute the MCM to health care workers and their families. Health care workers, considered critical infrastructure personnel, are then available to treat victims of a terrorist attack and maintain the level of medical support needed for a community. In addition to the benefits to patients needing care, communities may be more resilient if health care systems remain intact in the face of an attack. This prepositioning strategy also enhances equitable access to MCM by providing an alternative dispensing method for health care workers who will be expected to report to and stay at work during the course of the response to an attack, and who therefore will be unable to stand in line at PODs to receive MCM for themselves and their families. MCM caches in health care settings also could be designated to protect existing patients or residents of long-term care facilities from anthrax exposure, although these populations may be at lower risk for exposure.

Health care settings, especially hospitals and acute care facilities, generally are not well suited to serve as open PODs that dispense MCM to the public because serving this function likely would overwhelm the facility and distract from its essential function of providing health care during an emergency. Nonetheless, hospitals are viewed by many in their communities as a repository for essential resources for surviving a disaster or terrorist attack, and some hospitals could be overwhelmed with community members during a response to an anthrax attack. Communications and public education about where people should go to receive prophylaxis would be crucial in the event of such an attack (see Chapter 3).

Caches in Hospitals and Acute Care Medical Facilities

Hospitals and acute care medical facilities are part of the critical infrastructure for combating an anthrax attack and maintaining the health status of a community. Ensuring timely prophylaxis to this community, in particular, has a multiplicative effect: incapacitating one health care worker could negatively impact the care of dozens or more patients. If no other strategy for getting MCM to these facilities quickly is in place, prepositioning may be an appropriate strategy for protecting this infrastructure, particularly in high-risk areas.

The current emergency management standards of the Joint Commission do not specify that hospitals should have pharmaceutical caches for use in disasters (Live Process, 2011). However, many hospitals and health systems have developed pharmaceutical caches through the Hospital Preparedness Program (HPP), which was established in 2002 to enhance surge capacity and preparedness for public health emergencies among hospitals and health systems.2 HPP funding is provided to states, territories, and eligible counties, which in turn work with hospitals and health systems in their jurisdiction and pass along to them a portion of the funding for preparedness planning and exercising (ASPR, 2011a). During the first few years of the HPP, the development of pharmaceutical caches was a focus of the program, along with other capacity-building activities such as decontamination, development of bed surge capacity, and training for providers in diagnosing diseases caused by bioterrorism (ASPR, 2011b). Awardees were required to develop regional pharmaceutical caches containing a 3-day supply of antibiotics for hospital personnel, first responders associated with the hospitals, and their families (HRSA, 2004, 2005). These pharmaceutical caches did not necessarily have to be on-site at the hospitals; strategically placed regional caches also met the requirement. Although data on national prevalence are sparse, some major health systems maintain on-site pharmaceutical caches for staff in addition to state, regional, or local health care caches.3 Based on data from the HPP 2006 midyear progress report, 20 states reported that all their participating hospitals had pharmaceutical caches that could provide a 3-day supply to cover hospital personnel, associated first responders, and their families. Another 6 states reported that 90 to 99 percent of participating hospitals could meet this requirement (GAO, 2008).

After the first few years of the HPP, however, the focus shifted from capacity building to a capabilities-based approach. Hospitals now must demonstrate the capability to perform core response functions and no longer can meet requirements simply by purchasing equipment and supplies (ASPR, 2011b). Pharmaceutical caches became a level two subcapability, meaning that funding can be used for this purpose only if all capabilities designated as level one have been adequately addressed (ASPR, 2007, 2008, 2009, 2010, 2011c; HRSA, 2006). The grant guidance notes that most awardees should already have pharmaceutical caches because of the emphasis in earlier years, but a review of the first 5 years of the HPP notes, “Continued funding for hospitals is needed to pay for training of hospital staff, employment of hospital disaster coordinators, and maintenance or replacement of stockpiled supplies and pharmaceuticals purchased through the HPP” (Toner et al., 2009, p. 61). This report goes on to warn, “Although the [new] emphasis on coalition development is critical, it is clear that progress will be lost and individual hospitals will drop out of the HPP if they do not continue to receive some support for stockpiling, replenishing caches, and training” (Toner et al., 2009, p. 61). The new 2011–2012 effort to align the required capabilities of two major federal preparedness grants, the HPP and CDC's Public Health Emergency Preparedness (PHEP) Cooperative Agreement, is intended to promote coordination and efficient use of resources (Lurie, 2011). It should be noted that the ability to maintain hospital caches will be sensitive to the availability of ongoing funding.

Caches in Nonhospital Health Care Facilities

Antibiotics may also be cached in nonhospital health care facilities, such as community health centers, clinics, skilled nursing facilities, and subacute care facilities. Some of these facilities may be part of the same health network as larger hospitals with more fully developed caches, in which case hospital plans could be leveraged and extended. Prepositioning MCM for health care workers in a range of institutional facilities outside of hospitals and acute care facilities not only would help ensure that those workers could continue to care for patients and residents, but also would expand the pool of health care workers available to engage in a broader community response should an emergency overwhelm the internal surge capacity of hospitals. MCM caches in locations such as skilled nursing facilities also would provide protection for the patients and long-term residents, many of whom would be unable to stand in line at open PODs to receive their antibiotics.

Community health centers (CHCs) are somewhat different from skilled nursing facilities because they serve community members who are more likely to be out and about and, therefore, potentially at higher risk of anthrax exposure. CHCs have the ability to reach difficult-to-serve and vulnerable populations, including the medically underserved, people of low income and limited resources, people without insurance, migrant workers, and the homeless (Muccio, 2011). CHCs also may have the ability to tailor medication sheets and messaging to the multilingual populations they serve every day; they may have mobile assets to assist with further dissemination; and, like other health care facilities, they already have qualified, experienced staff to handle medications. However, CHCs may ultimately not be appropriate places to cache medications for use in open PODs serving their usual populations. First, CHCs and associated workers may be called upon to provide surge health care and would be distracted from this function by providing an open POD for the public. Second, CHCs may lack adequate storage space and security for the cache and may have concerns about the costs and time associated with maintaining it (Muccio, 2011).

Caches in Retail Pharmacies

As a routine component of the health care system, retail pharmacies are potentially appropriate sites for MCM caches. While incorporating them into the overall MCM dispensing system would require public-private coordination, discussed in the next section, they offer the benefit of trained staff and experience in storing and dispensing pharmaceuticals. Like hospitals, they can minimize costs associated with drug expiration by rotating cached antibiotics through regular use. However, caching antibiotics on-site does represent a variation from current pharmacy practice. Retail pharmacies usually stock medications only in the quantities needed to meet immediate needs, and they rely on pharmaceutical distributors and vendors to provide “just-in-time” inventory on a continuous basis. Additional costs also are associated with storing and maintaining excess inventory. In addition, the stock at a pharmacy likely would be in bulk, not unit-of-use packages, complicating the process of dispensing to the public. These factors make it unlikely that retail pharmacies would cache a sufficient quantity of antibiotics to supply an open POD. Instead, a retail pharmacy open POD would most likely dispense antibiotics supplied postevent by the SNS, state or local stockpiles, or vendors. On the other hand, the obstacles of stockpiling antibiotics to act as an open POD would be less of a challenge for pharmacies maintaining small caches of antibiotics just for pharmacists and other staff members and their families. The presence of such caches might encourage staff to report to work quickly and stay there throughout the response to a terrorist attack.

Public-Private Coordination and Workplace Caches

Private-sector entities may be interested in developing systems through which they could dispense antibiotics to employees and their families to help ensure their well-being, provide for continuity of business operations, and potentially lower insurance costs. In some cases, private-sector entities also have demonstrated willingness to assist in distributing and dispensing MCM for the general public (Lien et al., 2006). Many large private-sector companies have infrastructure and expertise for supply chain management and logistics, systems for communicating with employees, occupational health clinics and medical staff, and other resources that could be used to enhance dispensing capacity within their community during a time of crisis. By providing prophylaxis for employees and their families, private-sector entities also could alleviate the burden on the public health system and enable public health officials to focus more on dispensing to vulnerable populations that might not be reached by large employers. At the same time, however, potential private-sector partners face many barriers, including liability, cost, legal and regulatory issues, and the complexities of working across multiple jurisdictions during the development of MCM dispensing plans.

This section first considers private-sector participation in MCM distribution and dispensing in general, including current examples, advantages, and barriers. It then addresses the issue of the need for national-level guidance to inform public-private coordination on MCM prepositioning, distribution, and dispensing. Finally, the section examines the specific case of workplace caches, in which private-sector entities preposition MCM on-site. As discussed in more detail below, although many companies have expressed willingness to dispense SNS or state antibiotics via closed PODs, many have significant concerns about caching the MCM on-site.

Private-Sector Participation in MCM Distribution and Dispensing

Many private-sector entities played important roles in the response to the 2009 H1N1 influenza pandemic, demonstrating a willingness to partner with public health entities throughout the MCM distribution and dispensing system. The programs developed at that time greatly enhanced public-private coordination and demonstrated the expertise and capacity of the private sector, but they also brought to light some significant barriers to increased private-sector participation.

The SNS: Supply Chain Dashboard was developed in the fall of 2009 as part of the response to the 2009 H1N1 influenza pandemic to provide timely information about MCM in the commercial supply chain and public-sector stockpiles. This information assisted federal and state officials and other decision makers in responding to the pandemic (CDC, 2010b). Commercial partners submitted information each week on the available supply of MCM, their ability to fulfill orders, and upcoming production. Participants at a 2009 Institute of Medicine (IOM) workshop noted that this project involved an unprecedented degree of data sharing on the part of private-sector partners and a new level of coordination and communication among public and private stakeholders in addressing a public health emergency (IOM, 2010b). CDC plans to maintain the Dashboard as part of its Countermeasure Inventory Tracking program (CDC, 2011b).

Retail pharmacies played an important role in providing vaccine and, in some states, antiviral medications during the response to the 2009 H1N1 influenza pandemic (ASTHO, 2009; IOM, 2010a). Facilitated by the Association of State and Territorial Health Officials, many private and public entities came together quickly to create a framework for state and territorial health officials to partner with pharmacies to administer H1N1 vaccine. The organizations involved included the American Pharmacists Association, CDC, the National Alliance of State Pharmacy Association, the National Association of Chain Drug Stores, the National Association of County and City Health Officials, the National Community Pharmacists Association, the Office of the Assistant Secretary for Preparedness and Response (ASPR)/Department of Health and Human Services (HHS), DHS's Office of Health Affairs, and Pharmaceutical Research and Manufacturers of America's (PhRMA's) Rx Response (ASTHO, 2009). This model extends public health capacity to reach large numbers of people because privately employed pharmacists are a large workforce, pharmacies already have the staff and infrastructure to store and dispense MCM, and pharmacies generally are in well-known and accessible locations throughout communities. For these same reasons, this model also could be highly effective for dispensing antibiotics in response to an anthrax attack. However, it is critical that a framework for cooperation among public health officials and retail pharmacies be in place prior to an anthrax attack as the timeline of the response would not allow for this to be accomplished postevent.

Some private-sector entities have volunteered to serve as closed PODs that would dispense SNS or state MCM to employees and their families. The committee is unaware of data on how many private-sector entities have developed plans to serve as closed PODs after an anthrax attack, but this is a major component of the overall dispensing strategy in some jurisdictions. Speaking at a 2008 IOM workshop, Pamela Blackwell, Director of the Center for Emergency Preparedness and Response for the Cobb and Douglas Boards of Health, estimated that the closed PODs planned at that time for the metropolitan Atlanta area would serve enough people to reduce demand on the open POD system by 40 to 50 percent (IOM, 2008). At the same workshop, Teresa Bates of Tarrant County, Texas, reported that several large employers in the county have partnered with public health authorities to plan closed PODs (IOM, 2008). Some private-sector entities that plan to serve as closed PODs, dispensing only to employees and their families, also plan to provide volunteers to public health open PODs once the closed POD has completed dispensing (Buehler et al., 2006). A 2005 survey of private retailers suggested their willingness to serve as PODs. These respondents included many retail pharmacies—chain and independent pharmacies operating as stand-alone stores or through supermarkets—that, as noted earlier, already possess the physical space for staff training and dispensing and have an existing relationship with the public (Lien et al., 2006).

At the same time, participants representing the private sector at 2009 and 2010 IOM workshops cited several barriers that could discourage private-sector entities from becoming more involved. These included concerns about liability; compliance with federal and state laws and regulations; payment and reimbursement issues; communications; and, especially, working with multiple jurisdictions across the nation (IOM, 2010a,b).

Public-Private Coordination

No federal/national-level guidance currently exists on private-sector participation in an antibiotic prophylaxis campaign. Private-sector entities interested in prepositioning, distributing, and dispensing antibiotics must work directly with state and local public health authorities to coordinate on distribution and dispensing plans, and they must comply with varied state and local laws and regulations in addition to federal laws and regulations. Companies with multiple facilities nationwide must implement multiple policies and procedures instead of adopting a single corporate dispensing protocol. This fragmented approach has been cited as a significant challenge to increased participation by large national corporations, in presentations to this IOM committee and during previous workshops hosted by the IOM on MCM distribution and dispensing and on the 2009 H1N1 vaccination campaign (IOM, 2010a,b; Stargel, 2011; Turnbull, 2011).

Many states and localities already have developed their own memorandum of understanding (MOU) or memorandum of agreement (MOA) templates for private-sector entities interested in serving as a closed POD, specifying the respective roles, rights, and obligations of the public health agency and the private-sector entity. No central repository of existing MOUs or MOAs exists, and the committee is unaware of data on how many states and localities have developed them, but it believes that most states and many localities have done so.4

National guidance for public-private coordination in prepositioning, distribution, and dispensing would facilitate private-sector participation in these activities by promoting consistency across the nation. The federal government should convene state, local, and tribal governments and private-sector entities to develop such national guidance. The latter group should include representatives of businesses of different sizes, from different geographic locations, from both critical infrastructure and noncritical infrastructure industries, and from both health care and non–health care sectors. The federal government should also ensure that the plans developed in the national guidance include Public Readiness and Emergency Preparedness (PREP) Act coverage, without which private-sector entities are unlikely to participate. CDC should serve the convening role since it has primary responsibility for MCM distribution and dispensing at the federal level, has existing relationships with all public health authorities and many large private-sector entities, and already has been involved in developing public-private models for public health preparedness and response. This national guidance would be informed by the relationships already forged between the private sector and state and local agencies in response to real-life events (e.g., the 2009 H1N1 influenza pandemic).

Although the federal government may play a convening role, the guidance developed must ultimately be national–level guidance, and MOUs and MOAs must still be signed at the state level (even if the template is consistent across states). State and local public health authorities are responsible for dispensing MCM to the general public. Furthermore, most disasters are local rather than national, and therefore the declaration of a disaster will be at the state level. For example, there was no federal Stafford Act declaration of emergency for the 2001 anthrax attack.

The national guidance should address the range of roles that private-sector entities might play in the MCM distribution and dispensing system, including logistical support for distribution and dispensing, retail pharmacies dispensing to the general public, closed PODs in all kinds of workplaces, workplace caches of MCM, and private-sector rotation of state and local stockpile material to decrease replacement costs. Although the proposed national guidance (and the process for its development) would include consideration of prepositioning and workplace caches—the topic of this study—it is critical that these strategies be considered within the overall context of enhancing private-sector participation rather than being addressed in isolation. Box 4-2 outlines key components that might be included in national-level guidance for public-private coordination in prepositioning, distributing, and dispensing MCM.

Box Icon

BOX 4-2

Key Components of National Guidance for Public-Private Coordination in Prepositioning, Distribution, and Dispensing of Medical Countermeasures. The national guidance might include (but should not be limited to): mechanisms for sharing threat assessments (more...)

Across the various components of the national guidance, plans should leverage private-sector partners' existing systems and networks wherever possible. This would likely increase private-sector entities' willingness to participate since it would entail a lesser burden in terms of both cost and effort while also taking advantage of private-sector expertise and design efficiencies.

Finding 4-1: Private-sector entities may be interested in developing systems through which they can preposition, distribute, and dispense antibiotics to help ensure the safety of employees and their families and to provide for business continuity. Many large private-sector companies already have systems through which they communicate effectively with their employees, and they often have medical staff and other resources that could be used to enhance dispensing capacity within their communities during a time of crisis. However, potential private-sector partners face many barriers, including liability, cost, legal and regulatory issues, and the complexities of working across multiple jurisdictions during the development of MCM dispensing plans.

Recommendation 4-1: Develop national guidance for public-private coordination in the prepositioning, distribution, and dispensing of medical countermeasures. The Department of Health and Human Services should convene state, local, and tribal governments and private-sector organizations to develop national guidance that will facilitate and ensure consistency for public-private cooperation in the prepositioning, distribution, and dispensing of medical countermeasures and help leverage existing private-sector systems and networks.

Workplace Caches

As discussed above, workplace caches are one potential way in which private-sector entities could participate in MCM distribution and dispensing. Here the committee examines the specific strategy of private-sector prepositioning of MCM on-site in the workplace or storage of MCM on the company's behalf in a nearby pharmaceutical distributor warehouse.

Most existing plans to dispense antibiotics in workplaces via closed PODs rely on postevent supplying of the MCM by the SNS or a state or local stockpile. However, closed PODs could dispense MCM from several potential sources:

  • SNS or state or local stockpile (distributed postevent),
  • manufacturers or pharmaceutical distributors (distributed postevent),
  • a cache on-site at the workplace (prepositioned), or
  • dedicated caches at nearby pharmaceutical distributor warehouses (prepositioned).
Potential advantages

As noted earlier, closed PODs at workplaces could decrease the burden on public health open PODs regardless of the source of the MCM. Prepositioning MCM in on-site workplace caches or in dedicated caches in nearby pharmaceutical distributor warehouses also would alleviate the burden on the SNS or state or local RSS system. On-site caching may be particularly advantageous for employers that already have occupational health personnel on staff, mitigating many of the logistical and legal challenges involved (discussed in detail below). Especially in the case of large companies with many employees, closed PODs served by prepositioned caches could therefore alleviate the burden on the entire public health distribution and dispensing system and enhance the overall dispensing capacity in a jurisdiction while also potentially reducing costs by relying on private-sector efficiencies.

For employers, having an on-site MCM cache could minimize absenteeism due to incident-related concerns, allowing the company to maintain critical operations or to recover from a resulting business interruption more quickly. In a public health emergency, this capability is especially critical for hospitals, health care providers, and public and emergency services, as well as for critical infrastructure (e.g., public utilities). With on-site MCM caches, employers could provide prophylaxis to employees even if the public health distribution system became overwhelmed following a large-scale anthrax attack. However, additional costs and complications are associated with moving from a closed POD model with MCM supplied by the SNS or a state stockpile to a prepositioned workplace cache model, as described in more detail below.

Current prevalence of workplace caches

The committee is unaware of examples of private-sector companies that have developed workplace caches of MCM for their employees and families. Many companies have expressed willingness to dispense SNS or state MCM via closed PODs but have significant concerns about caching the MCM on-site. These concerns include storage space, liability, pharmacy laws, internal and external command and control, and replacement costs, as described below.

Companies have shown more interest in stockpiling antiviral medications, and in 2008, CDC issued guidance to advise employers that are interested in stockpiling antiviral medications for pandemic influenza (CDC, 2008a). In 2009, the H1N1 influenza pandemic greatly increased private-sector interest in MCM distribution and dispensing (IOM, 2010a). It is not clear, however, whether the systems developed to administer influenza vaccine would transfer directly to a response to threats such as anthrax because in the latter cases, MCM would have to be administered much more rapidly than would influenza vaccine.


As mentioned above, most existing workplace (closed POD) dispensing plans rely on the MCM being supplied by either the SNS or a state stockpile. Logistical arrangements are outlined in an MOA and include the responsibilities of state and local governmental health agencies and the closed POD sponsor. The MOA addresses liability issues, particularly those associated with the PREP Act and Food and Drug Administration (FDA) Emergency Use Authorization (EUA) requirements, as well as required risk communication and any reporting requirements. Prepositioning MCM in these types of sites would be relatively simple administratively. The relationship already exists, legal issues have been addressed, and roles and responsibilities have been defined. Prepositioning would, however, entail two additional logistical issues: secure and climate-appropriate storage, and MCM replacement upon expiration and disposal of the expired MCM.

Workplaces considering a prepositioning strategy face the challenge of finding space to house a stockpile and maintaining it under appropriate climatic conditions to ensure that the medication remains potent and to comply with prescription laws. The average dimensions of a pharmaceutical pallet are 48 by 48 inches (Missouri Department of Health and Senior Services, 2008); for workplaces with large numbers of employees, identifying physical space for storage could be an obstacle. Both antibiotics currently stockpiled for anthrax prophylaxis must be stored in a dry climate at approximately room temperature (ciprofloxacin must be stored below 86°F [FDA, 2011], while doxycycline can be stored at room temperature, at 68-84°F [NLM, 2008]). In addition, workplaces would need a permit from their state pharmacy regulatory agency to store MCM (or any other prescription medication) on-site (NABP, 2010).5 Medical professionals might not be necessary to staff a closed POD after an emergency since the emergency declaration could authorize nonmedical personnel to dispense MCM, but medical professionals would be needed to purchase MCM and store them in a workplace cache in advance of an emergency declaration. Businesses that employ medical professionals or that have occupational health programs may be particularly well suited to developing workplace caches since the required staffing already exists.

Appropriate security measures would be needed to safeguard the cached MCM and to ensure their secure transportation if storage and dispensing were carried out at different locations. Organizations also would have to plan for the safety of the personnel securing and dispensing the medications. If dedicated personnel were required for this task, that cost would have to be factored into the overall cost of maintaining a cache.

Private-sector entities would be responsible for monitoring the expiration dates of cached MCM, disposing of expired MCM, and replenishing the cache. These processes increase the logistical burden and costs of maintaining a cache. Currently, no mechanisms are available to defray the costs of replenishing expired stock, as only federal stockpiles are eligible for the SLEP (see Chapter 3) (Courtney et al., 2009). Unlike caches at health care locations, workplace caches cannot be rotated through routine health care delivery using a first-in/first-out approach to avoid expiration.

Legal considerations

Private-sector entities serving as closed PODs would have to consider many applicable federal and state laws and regulations. The Public Health Law Network has outlined the following legal issues relevant to private-sector entities that serve as closed PODs:

  • impact of an official declaration of emergency, disaster, or public health emergency;
  • MOU between a public health agency and an entity;
  • ownership of medical supplies;
  • medical personnel;
  • authorization to dispense medications;
  • EUAs;
  • liability;
  • workers' compensation;
  • privacy (Health Insurance Portability and Accountability Act [HIPAA] Privacy Rule and state privacy law compliance); and
  • reporting and documentation requirements (Public Health Law Network, 2011a).

In addition, employers caching medications on-site would have to consider some other legal issues, including compliance with state laws for pharmaceutical distribution, prescription, and storage; liability concerns associated specifically with the cached medications; liability if MCM were dispensed outside of a declared emergency; the possibility of the government's taking the MCM; and concerns about setting up an entitlement for employees. As noted above, laws and practices vary significantly from state to state, which greatly complicates the situation for large national companies.

The primary legal obstacle for non–health care entities is identifying and complying with state and federal laws governing prescription medications and the purchase of medications (CDC, 2008a). Private-sector entities seeking to cache antibiotics for anthrax would first have to identify someone licensed to have them and therefore allowed to purchase them. Companies that employ medical professionals, such as in an occupational health program, might be able to use those individuals to fill this role. Without such a licensed staffer, however, organizations would have to look to other mechanisms for obtaining MCM within the confines of the law. One remedy would be for the organization to contract with a physician organization. The physicians would write the prescriptions for the MCM to be stockpiled, and a mail-order pharmacy would deliver the MCM to the organization (Shulman, 2011).

Once a cache had been established, private entities or other workplaces would have to monitor the expiration date of the MCM and ensure that expired medications were disposed of and replaced. Federal laws regulate expiration dates placed on medication in its original packaging (usually in mass containers of hundreds of doses that must be separated to be dispensed), while states often establish the expiration of individual prescriptions at 1 year from the time the prescription is filled.6

Regardless of whether antibiotics were prepositioned or received from the SNS or a state stockpile, non–health care entities dispensing MCM would face certain liability issues. Under noncrisis conditions, dispensing prescription medications outside of the traditional health care system (at the workplace) exposes organizations to significant legal liability for any adverse events that might be experienced. During a federally declared emergency, however, the provisions of the PREP Act extend liability protections to all entities and individuals involved in the distribution and/or dispensing of approved MCM—including nongovernmental entities and persons (see Chapter 3 for further discussion; Public Health Law Network, 2011b). As noted in Chapter 3, however, some private-sector representatives continue to cite liability concerns as a barrier to increased participation in MCM distribution and dispensing plans (IOM, 2010b).

Entities that prepositioned antibiotics would face specific legal concerns that could increase their risk of liability compared with those acting as a postevent POD. First, if an organization were to dispense MCM prior to an approved EUA, it would be operating outside of its legal authority and the protection offered by the PREP Act. In contrast, SNS materiel is likely to be covered automatically by the PREP Act by the time it reaches PODs. Second, organizations could be liable for MCM cached on-site that were stolen and then used in a way that caused harm. This concern might be decreased by storing the medications securely and developing processes for access. Third, even during an emergency, PREP Act coverage does not limit an organization's liability if the MCM was stored improperly. Constant monitoring of cached materiel could pose both a logistical challenge and a financial burden. Fourth, in developing a workplace cache, employers could be perceived to be establishing a contract to provide employees with MCM after an attack.7 If they were unable to meet that obligation, perhaps because of extenuating circumstances arising from the aftermath of the attack, the organization could be held liable.

Finally, while there are no federal laws allowing or prohibiting public officials from reallocating private stockpiles during an emergency (known as “taking”), there is concern that the breadth of emergency powers granted to many governors might make public seizure of private stockpiles a reality (Gostin et al., 2002). This scenario was voiced as a barrier to private-sector prepositioning at the committee's public workshop (Turnbull, 2011). The PREP Act does afford private businesses some reassurance because it denies liability protection to assets that have been seized. The private and public sectors also can enshrine private ownership of private caches through independent legal agreements (MOUs or MOAs) that complement or supersede states' emergency powers; however, local governments may want to retain the right to take a private stockpile if the crisis should require doing so (Mathias, 2011). Public-private coordination in an overall dispensing system would decrease the incentive for governments to seize private stockpiles because those stockpiles would be seen as part of a single system, benefiting the entire population by off-loading demand on public PODs.

Caches in Other Non–Health Care Settings

Caches also could be established in community- and faith-based organizations and educational institutions.

Community- and Faith-Based Organizations

Service networks that serve vulnerable populations could play an important role in enhancing access to MCM for vulnerable populations, including people with low incomes and/or limited transportation, people with no or limited English proficiency, historically underserved ethnic/racial groups, people with disabilities, people who are homeless, and people who are homebound. These service groups include, for example, mutual association groups based on language or cultural commonalities; member organizations of the National Voluntary Organizations Active in Disaster; and service organizations such as Meals on Wheels, the Red Cross, local agencies on aging, and home health care services (Janis, 2011; Silver, 2011). Some vulnerable populations may not be best served by traditional public PODs because of issues of lack of trust of government services, mobility, and other demographic factors (Janis, 2011). However, community agencies could help provide better MCM access to the populations with whom they already interact by serving as PODs or by dispensing MCM via their usual service delivery systems.

Similarly, houses of worship often participate in the traditional MCM dispensing system by serving as open PODs and channels of reliable public health communication. The availability of staff and volunteers, status as a trusted place of gathering, and ability to accommodate the influx of a large number of people make houses of worship suitable to assist in dispensing MCM. In addition, houses of worship frequently are organizations with the ability to reach vulnerable groups such as non-English-speaking populations. Community groups that focus on disaster response, such as Community Emergency Response Teams (CERT), could also potentially play a role in dispensing MCM in their neighborhoods.

However, most community- and faith-based organizations lack the infrastructure and staff required to preposition MCM. These organizations may be better suited to serve as PODs or use other existing service delivery systems to dispense MCM delivered postevent. Some agencies—for example, food banks with their climate-controlled warehouses—may already have the capacity to provide appropriate storage conditions (Smith, 2011). However, these agencies would be unlikely to have medical staff permitted to purchase and store MCM.

Certain faith-based and other nonprofit human service organizations may have internal expertise in storing and dispensing medications and could be identified as potential prepositioning partners. These organizations might include, for example, home health care services and houses of worship with established health programs, such as faith community nursing, which promotes holistic and preventive care within faith communities. Prepositioning MCM with these agencies could alleviate the postattack burden on the traditional distribution system if the agencies were willing and able to meet the regulatory requirements of maintaining caches.

Educational Institutions

Schools have been suggested as logical venues for prepositioning, especially considering the need to address concerns that children have equal access to MCM (Anderson, 2011). Primary and secondary schools have both an infrastructure for and experience with interacting with children and their families, sometimes employ nurses, and often have the space required for MCM storage. Yet as large, trusted gathering places, schools usually are already involved in the local dispensing system as public (open) PODs, making it difficult to rely on them as closed PODs for students and staff as well.

Universities, on the other hand, might be better equipped to preposition and dispense MCM without the previous commitment to act as open PODs. They employ health care personnel and experienced staff, have the space for MCM storage, often are the place of residence for large numbers of students, and in some cases are the largest local employer (Turner, 2011).


Predispensing occurs when MCM, such as antibiotics, are stored by the intended users or by heads of households or other nonmedical caregivers for use by those in their care. Predispensing MCM is unique relative to other potential prepositioning strategies because it puts the MCM directly into the hands of the intended end-users. This introduces potential health risks to both individuals and the community that do not exist for prepositioning strategies such as forward-deployed and cached MCM. Predispensing MCM also involves a different set of logistical, communications, behavioral, and legal considerations compared with prepositioning strategies in which the MCM are not stored by the end-users.

The development of a strategy for predispensing MCM involves consideration of both function and form. Function refers to the role of the strategy within a jurisdiction's overall MCM dispensing plan. For example, predispensing could be used to dispense MCM broadly to the general public in a community, or it could be used only to target specific subpopulations or individuals, such as those who lack timely access to MCM through other mechanisms. Form refers to the specific manner in which the MCM is predispensed, including the following:

  • Personal stockpile: MCM that is dispensed to individuals pre-event via normal prescribing routes for use during a public health emergency.
  • MedKit: A medical kit containing prescription pharmaceuticals that is dispensed pre-event to families or individuals for use only as directed during a public health emergency.

    EUA MedKit: A medical kit approved by the FDA under its EUA for off-label use.

    FDA-approved MedKit: A medical kit approved by the FDA for labeling and use as a predispensed MCM.

An over-the-counter MedKit is a theoretical possibility, but the committee did not consider this to be a feasible option since the FDA has never approved any antibiotics for over-the-counter use, and this strategy also would run counter to current public health efforts to restrict the widespread use of antibiotics, as discussed in greater detail below.

This section of the chapter is organized primarily by function: it first examines predispensing MCM to the general public in a community and then examines predispensing to targeted subpopulations within a community. This latter strategy is illustrated using two examples: certain first responders and critical infrastructure workers and their families, and selected patients. For each function, the committee discusses potential benefits; concerns about inappropriate use; concerns about the flexibility and adaptability of the strategy; storage and stability issues; and practical considerations such as logistical burdens, legal issues, and communications. The discussion of each function also includes the potential impact of the form in which the MCM is predispensed. First, however, an overview of the available evidence on predispensing strategies is presented.

Overview of Available Evidence

The evidence base available for assessing the use of predispensing strategies is limited. The predispensing of antibiotics for anthrax has been tested in one pilot study in St. Louis, Missouri, and has been implemented for a limited group of postal carriers who volunteered for the Minneapolis-St. Paul pilot of the postal model. These examples are summarized below and provide some insight into predispensing of antibiotics for specific subpopulations; however, significant limitations hinder generalization to the general public and to circumstances beyond those in the St. Louis study and the Minneapolis-St. Paul pilot program. Predispensing of antibiotics has never been tested in an actual emergency or tried with the general population.

The committee also considered evidence from other potential models, including the misuse of antibiotics prescribed during routine medical care; the predispensing of potassium iodide (KI) to those living within 10 miles of a nuclear power plant; and the general public's response to crises in which MCM are indicated only for a narrowly defined potentially exposed population, such as during the 2001 U.S. anthrax attack and the 2011 nuclear accident in Fukushima, Japan.

Some examples of predispensing MCM exist in other countries, such as a previous program to preposition antidotes for nerve agents in homes in Israel (suspended because of a lack of cost-effectiveness) (Stoil, 2010). Because of the different political and cultural environments and the differences in health care systems and regulations, however, the committee did not think these examples would provide reliable evidence to inform the use of predispensing in the United States.

The sources of evidence considered by the committee are introduced here, with brief discussion of how each source informs (or does not inform) the evaluation of predispensing strategies for anthrax, in both function and form. The specific findings from each source are discussed throughout the following sections on predispensing for the general public and for targeted subpopulations.

Misuse of Antibiotics for Routine Medical Care

Antibiotics are commonly prescribed during routine medical care, and their misuse also is common, including failing to complete the recommended regimen, skipping doses, and reusing leftover doses (e.g., Kardas et al., 2005). The committee considered available data on misuse of routinely prescribed antibiotics to inform conclusions about the likelihood that the general public will use predispensed antibiotics—intended for use during an anthrax attack—outside of a declared emergency. The committee found no data on antibiotic misuse that would inform predispensing for particular targeted subpopulations. In terms of the form of predispensing, this example aligns most closely with personal stockpiling of antibiotics for anthrax, since both involve standard prescription vials. However, provision of a regular prescription bottle specifically as a predispensing method for anthrax has never been studied, and no evidence is available to help determine whether antibiotics dispensed for protection against anthrax would be treated the same as those dispensed during routine medical care. In the absence of research examining this question, the committee judges that the two situations would likely show comparable rates of misuse.

MedKit Pilot Study: St. Louis, Missouri

In 2006, the Missouri Department of Health and Senior Services, in collaboration with CDC, placed prototype MedKits containing a blister-packed 5-day supply of doxycycline in more than 4,000 homes in St. Louis (CDC, 2007b, 2008b). The study population comprised three cohorts: clients and some employees of a community health clinic, employees from 10 major corporations, and first responders. To ensure compliance with state and federal regulations, the prototype MedKit was classified as an Investigational New Drug (IND), and the prospective pilot study was conducted under an IND protocol. The pilot study was the first effort to test the ability of households to store and maintain the MedKits properly, including saving them for emergency use only, and to assess attitudes and perceptions regarding the kits and factors that might influence how participants maintained or used them. Limitations of the study that were discussed at a 2009 IOM workshop included that it did not test whether participants were able to follow the enclosed instructions for preparing and using the antibiotics accurately and safely (a general challenge for any dispensing mechanism), nor did the study test how the medication was impacted by actual storage conditions in participants' households (a significant challenge for predispensing as therapeutic effectiveness can be affected by improper storage) (IOM, 2010b).

Households were randomly assigned to follow-up after 2, 4, or 8 months, at which point an exit interview was conducted and the MedKit was collected. A monetary incentive was offered to participating households, consisting of a $25 gift card provided at the time of the initial interview and another $25 gift card upon completion of the follow-up interview.8

Participants were not necessarily representative of the general population of St. Louis or of the United States (e.g., with regard to level of education or incentive to participate). However, the three study cohorts did have varied characteristics:

Clinic Cohort generally had lower levels of educational achievement, employment, and household income than the other two cohorts, was more likely to be African American, and was less likely to have health insurance coverage. The Business Cohort respondents were two-thirds female, and one[-]third reported an annual household income of greater than $80,000. They also accounted for 50% of the study population's graduate-level education. The First Responder Cohort respondents were predominately male (67%) and the majority were married (68.5%). First responders were also more likely to have health insurance coverage (96.9%). (CDC, 2008b, p. 6)

Although the study participants were not representative of the general public, these variations in their characteristics—for example, a cohort with a lower level of education than average and two cohorts with higher education than average—mean that this study could potentially provide some sense of the range of behaviors that might be expected if predispensing were implemented for the general public. Given the sparseness of available data on predispensing, the committee carefully considered the results of this study but ultimately concluded that it has limited utility as a model for predispensing for the general public because some design features could have biased participants toward greater adherence to instructions than would be expected outside of the study environment. Specifically, the study offered a financial incentive for participation, there was a selection bias toward those interested in volunteering, the follow-up periods were short, and the level of supervision and screening might not be possible if the strategy were used on a much larger scale over a longer period of time.

The committee did, however, consider the evidence from the first responder cohort to be useful as a model for predispensing in targeted sub8 populations. In this case, (1) the study population was more aligned with the subpopulation under consideration, (2) the level of supervision and screening provided in the study could similarly be provided in the first responders' workplaces, and (3) the financial incentive to participate in the study could have an effect comparable to that of an employer directive to adhere to instructions. Concerns about selection bias still do somewhat limit the applicability of this evidence, however.

With regard to the form of predispensing, this model informs the use of a MedKit (either through an EUA or FDA-approved) in terms of the range of potential rates of inappropriate use. This model also aligns most closely with the committee's definition of an EUA MedKit in terms of costs.

MedKits for Postal Workers in Minneapolis-St. Paul

As discussed in Chapter 3, postal carriers who volunteered to participate in the Minneapolis-St. Paul pilot of the postal model for MCM distribution were provided with MedKits to keep in their homes, containing sufficient quantities of antibiotic for themselves and their families. The MedKits would help ensure that postal carriers would be protected as they delivered MCM to the community. The provision of the MedKits to volunteers was a condition of participation negotiated by the postal workers' union and the U.S. Postal Service (USPS) (IOM, 2010b). The MedKits contained a short-term supply of the antibiotic doxycycline; postal carriers with a contraindication for doxycycline (e.g., an allergy) could not participate in the pilot program. Under the program, individual MedKits also were kept in the participants' workplaces for use should an attack occur when the postal carriers were already at work.

To ensure compliance with state and federal regulations, an EUA was obtained to authorize the distribution of the MedKits—referred to as a “household antibiotic kit” in the EUA (FDA, 2009). According to the conditions of the EUA, the USPS must survey participants every 6 months regarding the status of their MedKits (e.g., storage conditions, expiration of antibiotic, use of antibiotic). The USPS also is responsible for collecting expired antibiotic and turning it over to the designated public health authority for disposal and accountability record keeping. Upon termination of the EUA, the USPS must collect all MedKits and turn them over to public health authorities. This program did not assess participants' ability and willingness to take the MCM as instructed.

As with the household MedKit pilot study in St. Louis, the participants in the Minneapolis-St. Paul pilot are not necessarily representative of the general population. Volunteers were generally white, male, older, highly educated, and likely to comply with guidance (Griffith, 2011). In addition, they volunteered to participate, and the MedKit program was developed in exchange for postal workers' agreement to participate in the pilot of the postal model for dispensing MCM to homes throughout the community. It could be argued that the families of postal workers can be considered representatives of the “general public” because, for example, they may not necessarily have the higher education levels of the family members who work for the USPS. However, postal workers' family members likely have a financial and social incentive to comply with instructions of the workers' employer regarding adherence and saving the kit for emergency use. Therefore, the committee judged that the evidence from the postal pilot informs primarily predispensing to targeted subpopulations.

Examples Involving Potassium Iodide

The only example of widespread MCM prepositioning in homes that has been employed in the United States is the dispensing of KI to those living within 10 miles of a nuclear power plant (Box 4-3). This example has limited use as a model for predispensing of antibiotics because of key differences between KI and antibiotics. First, there is no reason for the public to take KI other than for protection following a radiological/nuclear emergency, while some people could be tempted to take predispensed antibiotics to treat a perceived infection. Second, KI is not a prescription medication, meaning that it is not subject to the same legal and regulatory requirements as antibiotics. And third, KI is a stable salt, unlike antibiotics, which degrade and require regular collection and replacement potentially every year (in light of current prescription law), although the packaging does still have a printed expiration date as required by the FDA.

Box Icon

BOX 4-3

Prepositioned Potassium Iodide for People Residing Near Nuclear Power Plants. Potassium iodide (KI, provided in tablet form) prevents the thyroid gland from absorbing radioactive iodine that can cause cancer and death. Several thousand people across the (more...)

2001 U.S. Anthrax Attack and 2011 Japanese Nuclear Crisis

Some useful insights into social behavior may be gleaned from the public's response during the 2001 U.S. anthrax attack and the reaction in the United States following the nuclear accident in Fukushima, Japan, in March 2011 (for an overview of the former, see Jernigan et al., 2002, and of the latter, see IAEA, 2011).

Two research groups have reported that during and after the 2001 U.S. anthrax attack, there was a significant increase in patient requests for anthrax-related antibiotics in areas of the country not affected by the attack (i.e., areas outside of New York; Washington, DC; and Florida) (Belongia et al., 2005; M'ikanatha et al., 2005). The physicians surveyed reported not only an increase in patient requests but also an increase in their own prescribing of such antibiotics as a result of these requests. M'ikanatha and colleagues (2005, p. 1) conclude that “public fears may lead to a high demand for antibiotic prophylaxis during bioterrorism events.”

Likewise, the nuclear meltdown in Fukushima in the aftermath of the 2011 Japanese earthquake engendered some panic in the United States, despite messaging from the federal government that no risk of harmful radiation exposure existed to U.S. residents; some Americans still perceived the risk of harmful radiation exposure to be significant. The media reported that pharmacies around the country sold out of KI, that all three manufacturers and suppliers of FDA-approved KI treatments sold out just a few days after the earthquake and tsunami hit Japan, and that a surge in prices occurred for KI products sold online (Aleccia, 2011; Lazaruk, 2011). Reports also cited customers purchasing what they thought were KI pills, only to find later that the product was fake or of an inadequate dosage (Lazaruk, 2011). Unfortunately, no research has been done on the breadth of public concern and the extent to which the panic was fanned and exaggerated by the media. The only data the committee could find on how many people actually took KI in the United States after the meltdown in Japan come from an MSNBC.com article, which stated that the American Association of Poison Control Centers said seven adverse reactions to KI had been reported among its 57 poison control centers across the United States; two of these reactions were reported to be serious, including vomiting, racing heart rates, and vertigo (Aleccia, 2011).

Despite the extremely limited evidence, the committee considered these examples in its deliberations on predispensing of anthrax antibiotics for the general public because they are the only source of evidence that provides insight into how people might behave with predispensed MCM during an actual emergency. With regard to the form of predispensed antibiotics, however, KI does not align specifically with either personal stockpiles or MedKits since it does not require a prescription.

Table 4-2 provides an overview of the available evidence relevant to two functions of predispensing (for the general public and targeted subpopulations) and several potential forms of predispensed antibiotics. The table clearly shows the gaps in the available evidence. In particular, it highlights the complete absence of evidence on personal stockpiling and the lack of evidence with which to compare and select among the potential forms of predispensing: personal stockpiles and MedKits—EUA, IND, and FDA-approved. The results of these studies and the implications of these evidence gaps are discussed throughout the remainder of this chapter.

TABLE 4-2. Overview of Available Evidence to Inform Predispensing Strategies.


Overview of Available Evidence to Inform Predispensing Strategies.

Predispensing to the General Public

Predispensing has the potential to significantly benefit members of the general public after an anthrax attack by reducing the time to prophylaxis for those individuals possessing the predispensed antibiotics and potentially alleviating the burden on the public PODs, which in turn would reduce the time to prophylaxis for individuals receiving their antibiotics from those sites. These potential health benefits and the costs associated with this strategy are discussed below; a framework for assessing the trade-offs between these benefits and costs is presented in Chapter 5. As described below, however, the decision to employ predispensing as a general public health strategy also should take into account the significant concerns about potential health risks, the inflexibility of this strategy, and associated practical burdens.

The major determinants of the magnitude of the risks associated with predispensed antibiotics are the maximal acceptable delay in prophylaxis following exposure and the risk of an attack. With a shorter acceptable time to initiate prophylaxis and a higher risk of attack, the risks of adverse events and inappropriate use become more tolerable. With a longer acceptable time to initiate prophylaxis and a lower risk of attack, the risks of adverse reactions and inappropriate use become less tolerable. Likewise, as discussed further below, the less well defined the exposed population, the more people will self-treat inappropriately, thus increasing that risk.

As discussed in Chapter 2, the available scientific evidence indicates that dispensing of MCM should occur within approximately 4 days of the time of exposure. If an attack were detected within the first 48 hours or so, jurisdictions might be able to use other dispensing strategies to reach the exposed population without incurring the potential health risks and practical burdens associated with predispensing (discussed below). However, if the window to treat were briefer, communities might determine that the potential health benefits associated with predispensing would outweigh the potential health risks, costs, and practical burdens. Chapter 2 raises the possibility of a shortened time window for effective prophylaxis delivery due to the uncertainties inherent in the available scientific data on the incubation period for anthrax and the concern that detection of an anthrax attack could be delayed. In addition, the committee acknowledges that risk data could be available at the classified level that would indicate a shorter time frame for prophylaxis.

This section examines issues related to predispensing MCM to the general public: concerns about inappropriate use; storage and disposal issues; concerns about the flexibility of the strategy; and practical considerations such as communications, behavior and adherence, logistics, and legal issues. The discussion indicates how the form in which the MCM is predispensed would impact these issues.

Inappropriate Use

Safety is a primary concern associated with predispensing of antibiotics, particularly if predispensing targets the general public. This concern includes the potential for inappropriate use in routine settings (e.g., using the antibiotics to treat a cold) and the potential for widespread inappropriate use in response to a distant anthrax attack, a false alarm caused by a nonanthrax white-powder event, or some other public health emergency for which antibiotics are not indicated. In her testimony to the IOM committee, Nadine Shehab of CDC outlined several potential areas of concern: misuse, which could result in adverse reactions and increased antibiotic resistance; unintentional ingestion; dosing and dose delivery; drug interactions; stability and storage; and adherence (Shehab, 2011).

This section first outlines what evidence can be brought to bear to estimate the prevalence of misuse if antibiotics were predispensed among the general public, including a discussion of the potential impact of the MedKit's special packaging and emergency-specific instructions. The section then outlines the anticipated consequences of inappropriate use, including adverse reactions and the potential to contribute to increased community and antibiotic resistance.

Estimated prevalence

A meta-analysis of antibiotic misuse found that in North America, mean adherence to antibiotic regimens was 57.4 percent (95 percent confidence interval [CI]: 44.0-70.8 percent); failure to adhere included not completing the entire course, missing doses, and reusing leftover doses (Kardas et al., 2005). A survey of more than 1,300 patients conducted in 2001 found that 17 percent had “self-prescribed” antibiotics left over from another illness (Richman et al., 2001). Most of the latter patients responded that they had used the antibiotics to treat a cough or sore throat without consulting a physician. This body of evidence suggests that the rate of misuse of predispensed antibiotics is likely to be high.

The example of KI use in the United States after the Japanese nuclear accident demonstrates the potential for misuse during a perceived emergency even when there is little or no risk of exposure. As noted earlier, in addition to the rush to purchase KI, limited evidence suggests some people actually took it, with seven adverse reactions being reported. Although it cannot be confirmed, these adverse reactions likely were related to KI prophylaxis among individuals concerned about the Japanese accident since KI is not a recommended treatment for any other indication. The data on inappropriate use of KI are extremely limited; this is the only example of MCM use during an actual emergency. Given that the emergency occurred on the other side of the world, it appears likely that misuse of predispensed antibiotics would rise across the nation following a localized anthrax attack somewhere within the United States.

In contrast to the above two sources of evidence, misuse among participants in the St. Louis pilot study was very infrequent (CDC, 2007b, 2008b). There were 4,076 households in the final analysis:

  • 97 percent returned their MedKit (blister packs intact, although some outer bags had been opened out of curiosity);
  • 3 percent could not locate the kit (or refused to return it, in the case of five households); and
  • 0.1 percent of households (four total) admitted to having used the medication.

The clinic cohort was more likely to report having used the MedKit (4 of 1,443 households, 0.3 percent) than the business and first responder cohorts, which reported no usage (out of 1,077 and 1,556 households, respectively). No analysis was presented as to whether this difference among groups was statistically significant. Although the overall rate of having taken the medication was low, the differences between the study conditions and the conditions under which MCM would be predispensed broadly to the general population make it difficult to extrapolate from these results to the expected rate of misuse in a broad predispensing strategy for the general public. As described above, the financial incentive to participate, selection bias, and the short follow-up periods limit the conclusions that can be drawn from this study to inform predispensing of MCM to the general public in a community.

Similar rates of return without opening were seen in the 6-month survey of participants in the Minneapolis-St. Paul pilot postal program. The first status check survey, conducted 6 months after initiation of the pilot program, found that the vast majority of volunteers had maintained their kit unopened (Griffith, 2011). One year into the program, 367 of the 377 kits held by active volunteers were returned, unopened, for scheduled replacement. As discussed above, however, the committee found that the postal workers could not be considered the “general public,” and it would be a stretch to consider their families as such. Because of these limitations, the committee does not view this pilot program as compelling evidence in favor of dispensing MedKits for the general population.

The St. Louis MedKit study and the MedKit component of the Minneapolis-St. Paul pilot program suggest that misuse of predispensed MCM would not be highly prevalent, but these examples have significant limitations, as described above. The overall prevalence of misuse of antibiotics and the recent experience with KI in the United States cast doubt on the likelihood of the general population's adhering to the requirements for a safe and effective home storage program for MCM, and therefore increase concerns about adverse outcomes and cost-effectiveness (cost-effectiveness is discussed in detail in Chapter 5).

Impact of form of predispensing

It is not known whether the relatively low rates of misuse in the St. Louis study and the postal pilot are attributable to the special packaging used in MedKits or to other study features, such as the financial or other incentives to participate or the short follow-up period. Certainly the intent of the MedKit packaging is to decrease misuse, but no direct evidence exists linking low misuse rates to the form of packaging (e.g., a comparison of misuse rates from two cohorts, one of which is given a routine antibiotic prescription bottle and the other a MedKit). The most extensive body of evidence suggesting a high rate of misuse (of antibiotics generally) is aligned more closely with the committee's definition of a personal stockpile. However, no evidence exists as to whether antibiotics dispensed specifically for protection against anthrax would be treated in the same way as those dispensed during routine medical care, although it appears possible that the two situations would be comparable. These uncertainties call for a larger, more comprehensive study to test the impact of the form of predispensed MCM on rates of inappropriate use and to better test the feasibility of predispensing by omitting features that could artificially increase the rate of adherence to instructions, such as financial incentives for participation.

Finding 4-2: The most extensive body of relevant evidence (statistics on the misuse of antibiotics prescribed for routine medical care) suggests that inappropriate use would be high if predispensing were implemented broadly for the general public. There are no appropriate data to bring to bear on the question of whether the rate of inappropriate use of MedKits (as currently designed) would be lower than that for personal stockpiles. However, future studies may be able to demonstrate that special packaging for MedKits (or similar products) could decrease the rate of inappropriate use.9

Adverse events

All antibiotics are associated with significant rates of adverse drug events (ADE), even when used appropriately. FDA-approved package inserts for most prescription medications list many adverse reactions.10 In a systematic review of 24 doxycycline clinical trials (N = 3,833), the range of incidence of ADEs was 0 to 61 percent (Smith and Leyden, 2005). Although many of these ADEs were relatively minor (usually gastrointestinal), antibiotics also are associated with ADEs severe enough to result in an emergency department visit; indeed, 7 of the top 15 medications implicated in emergency department visits for ADEs in 2004–2005 were antibiotics (Budnitz et al., 2006). Data from the National Electronic Injury Surveillance System-Cooperative Adverse Drug Event Surveillance project, 2004–2006, showed that ADEs from antibiotics accounted for nearly one in five emergency department visits for ADEs, corresponding to approximately 140,000 visits annually in the United States (95 percent CI: 116,506-168,504) (Shehab et al., 2008). Although these ADEs were severe enough to lead people to visit the emergency department, 94 percent of the visits did not require admission (Shehab et al., 2008). Of the antibiotics indicated for the treatment and prevention of anthrax, doxycycline was associated with an estimated 5.8 (95 percent CI: 3.9-7.7) annual emergency department visits per 10,000 outpatient prescription visits; ciprofloxacin with 6.4 visits (95 percent CI: 4.5-8.4); and amoxicillin and penicillin with 15.5 visits (95 percent CI: 12.3-18.7) (Shehab et al., 2008).

Of 5,343 people who took at least one dose of antibiotics following the 2001 anthrax attack, 57 percent (N = 3,032) reported ADEs, but hospitalizations and severe ADEs were rare (Shepard et al., 2002). Of the 2,631 persons (49 percent) who failed to complete the recommended 60-day course of antibiotics, 43 percent indicated that the reason for discontinuation was ADEs.

Similar rates of ADEs could be expected for those who misused predispensed MCM. Broader and less targeted predispensing of MCM would likely increase the numbers of ADEs. These rates of ADEs would likely be acceptable following an anthrax attack, when the potential benefits of the antibiotics would outweigh the potential risks, but might not be acceptable at other times.

Contraindications and warnings

Some people are allergic to doxycycline and other antibiotics; of the estimated 140,000 emergency department visits for ADEs attributed to antibiotics, approximately 80 percent were due to an allergic reaction (Shehab et al., 2008). Doxycycline also can cause permanent tooth discoloration if taken during tooth development, and therefore it is not recommended during the last half of pregnancy or for infants or children under the age of 8 (Pfizer, 2011). Widespread availability of doxycycline in homes, combined with some evidence that family members often “share” prescription antibiotics, could conceivably increase the likelihood that a pregnant women or child would take doxycycline (Larson et al., 2003).

Antibiotic resistance

Infectious disease experts have expressed concern that misuse of antibiotics in home MedKits could contribute to the growing problem of antibiotic-resistant microorganisms (IDSA, 2008; Navas, 2002; NBSB, 2008). Again, this issue would be of particular concern if predispensing were carried out on a large scale for the general public instead of for targeted subpopulations since the potential for large-scale misuse would be greater. Public health experts view antibiotic resistance as a major public health concern. The concept of predispensing runs counter to major educational efforts by CDC, the Alliance for the Prudent Use of Antibiotics (APUA), and others regarding the rational and limited use of antibiotics. CDC's national Get Smart campaign aims to reduce antimicrobial resistance by promoting more appropriate antibiotic use (CDC, 2010c). Comparable campaigns are conducted at the state level (e.g., Get Smart Colorado, 2011). APUA has similar aims (APUA, 2010).

Overuse of antibiotics not only increases selective pressure for the development of antibiotic-resistant organisms in the community but also leads to carriage of antibiotic-resistant organisms in individual patients. An additional concern is that selection for resistance to one antibiotic can result in resistance to multiple antibiotics because genes mediating resistance often travel together on transmissible genetic elements, such as plasmids or transposons (Bennett, 2008). Therefore, overuse of doxycycline could also lead to resistance to other, more commonly used antibiotics, such as penicillin, cephalosporins, and fluoroquinolones (e.g., ciprofloxacin). Again, impact depends on the extent of use: if an antibiotic is used broadly, the potential for resistance is greater than if use is more targeted and restricted. It is unclear, however, whether use of antibiotics predispensed as MCM would be a significant contributor to resistance relative to other known factors, such as prescribing for nonbacterial infections, failure to adhere to dosage and length-of-administration instructions, and the use of antimicrobials in animal feed (Levy, 2001; Rambhia and Gronvall, 2009).

Storage and Disposal Issues

Storage and disposal issues include stability and storage conditions, the potential for unintended ingestion, and safe disposal.

Stability and storage conditions

Prepositioning antibiotics in homes raises a number of concerns about storage conditions. First, are the products stable under home conditions? Most expiration testing is conducted in a dark, cool, dry place, but many people store their medications in a warm, moist bathroom. In the St. Louis MedKit study, 56.2 percent of participants stored the kit in a bedroom, 75.4 percent stored it in a closet, and 99.4 percent stored it out of the reach of children and pets. In this study, participants were given careful instructions to store the MedKits in a cool, dark, dry place out of the reach of children and pets; it appears likely that if predispensing were implemented broadly for the general public, storage conditions might more closely approximate those found for most medications. The St. Louis study did not test the impact on the medication of the storage conditions in which participants kept their MedKits (IOM, 2010b).

Having degraded drugs would give the individuals storing them a false sense of security. If the MCM were replaced every year, they would likely be stable enough to provide adequate protection against anthrax, regardless of the storage conditions (Injac et al., 2007). However, this strategy would impose substantial cost and logistical burdens (cost issues are discussed in Chapter 5). Finally, there are concerns that the stored MCM could be lost or stolen.

Potential for unintended ingestion

If predispensed MCM were not stored out of the reach of children, the potential for unintentional/unsupervised ingestion by children would arise. Poison control centers receive approximately 1.5 million calls annually for pediatric overexposures, the majority of which are through ingestion (Bronstein et al., 2007). Approximately 58,000 emergency department visits annually are for unsupervised medication ingestion by children, and 1 of every 180 visits made annually by 2-year-olds to an emergency department is for a medication overdose (including unsupervised ingestion) (Schillie et al., 2009; Shehab, 2011).

Safe disposal of antibiotics

Expired antibiotics stored at home must be disposed of safely. Improperly disposed antibiotics and other drugs in the normal waste stream reach water supplies (Karthikeyan and Meyer, 2005). Expired antibiotics in a government-sponsored predispensing program would need to be collected and disposed of by a company with appropriate expertise, again adding to logistical and cost burdens. If predispensed MCM were available solely through an individual purchase and maintain system, they likely would frequently be stored long past their expiration date or disposed of improperly.

Impact of form of predispensing

No data are available comparing how individuals would store and dispose of MedKits versus personal stockpiles and the effect on the stability of the MCM. The committee judges that the use of MedKits might encourage better storage conditions since the kits could be designed using a dark box to protect the antibiotics from light and with large-font instructions about avoiding storage in a bathroom. Rates of unintended ingestion would likely be similar since MedKits would be sealed, while standard prescription vials have child-safety features. The concerns and burdens associated with safe disposal would be similar unless a special dispensation were obtained to allow an extended expiration date for MedKits relative to the usual state prescription law.


In addition to misuse (discussed above) and cost-effectiveness (discussed in detail in Chapter 5), the primary concern associated with predispensing to the general public is the strategy's lack of flexibility. There are several specific concerns, described below: lack of ability to quickly provide alternative MCM; access to household stockpiles during an attack; dosing, dose delivery, and screening for contraindications and drug interactions; and inability to serve multiple purposes relative to PODs. Flexibility is examined in this section as it applies to predispensing strategies in isolation. The committee recognizes that if predispensing is used in conjunction with more flexible strategies (e.g., open PODs), the impact of potential challenges can be mitigated. As discussed in Chapter 5, however, this approach is costly because it requires the establishment and maintenance of multiple dispensing mechanisms.

One of the greatest concerns with predispensing strategies targeting the general public is the possibility that the anthrax used in an attack could be resistant to the predispensed antibiotic. This is a major concern for the entire dispensing enterprise, but particularly for predispensing strategies because, unlike the use of PODs, these strategies involve no infrastructure that could be used to rapidly dispense a different MCM, the communication challenges would be immense, and the loss of public trust would likely be great. As currently conceptualized, predispensing would involve a single antibiotic, such as doxycycline. One conceivable way to mitigate the concern about drug-resistant anthrax would be to develop a predispensing model involving multiple MCM. After an attack occurred, public health authorities would give instructions such as “take the blue pill.” This strategy would be somewhat similar to the current Australian Flying Doctors model, in which a series of common medications, including antibiotics, are prepositioned in medical chests in rural towns throughout Australia. When residents become ill, they call a hotline to speak with a health care professional and are told which medications to take from the chest and how often (RFDS, 2011a,b). However, safety risks increase with multidrug models, and cost would increase as well since this strategy would likely entail adding medications beyond the current least expensive antibiotics, ciprofloxacin and doxycycline.

Another concern associated with predispensing antibiotics to be stored in the home is that people may not be at home or may not be able to get there in the event of a bioterrorism attack. This possibility degrades the presumed reduction in time to prophylaxis that is the driving force behind predispensing strategies. Moreover, while any dispensing strategy carries the risk that exposed individuals will commute outside of areas in which they will be able to receive timely prophylaxis, predispensing MCM may convey a false presumption of lower demand at public PODs, which will not be the case if many or most individuals are unable to reach their MedKits. As discussed above, postal carriers who volunteered to participate in the Minneapolis-St. Paul postal model pilot were provided with two MedKits: one to keep in their home and one for their workplace. Although providing more flexibility to adapt to the situation, this solution decreases cost-effectiveness and increases waste and disposal needs.

Special dosing and forms of dose delivery may be required for children, pregnant and lactating women, older adults, people with renal impairment, and adults who cannot swallow tablets. Concern has been raised about the lack of evidence on palatability and the ability of parents and other adults to follow instructions for preparing child doses (NBSB, 2008; Robbins, 2011). The committee concludes, however, that (unlike with the issue of antibiotic-resistant anthrax raised above) this is a surmountable issue. First, challenges of accurate dosing are common to all dispensing strategies and are not unique to predispensing. Deviations from proper dosage for children are minimized through a combination of increasing parents' health literacy, providing parents with easy-to-use dosing instruments (e.g., oral syringes versus cups), and utilizing color coding (Frush et al., 2004; Yin et al., 2010). Another concern regards screening for contraindications and drug interactions. Although screening could be performed when the antibiotic was predispensed, individuals' health situations could change; for example, women could become pregnant. The committee also concluded that this issue would be surmountable by using clear instructions in the packaging. Nevertheless, the committee notes that it may be easier to address these issues in a POD than in a predispensing strategy.

Lastly, POD infrastructure can be used for many public health functions other than postexposure prophylaxis against anthrax, such as influenza vaccination and postexposure prophylaxis following exposure to hepatitis. In contrast, predispensing likely would not serve multiple purposes, which suggests that PODs may be more cost-effective and a better use of planning resources.

The form of predispensing does not impact concerns about lack of flexibility related to antibiotic-resistant anthrax, lack of access to MCM stored at home during an attack, and inability to serve multiple purposes. MedKits may be able to mitigate some of the concerns related to dosing by, for example, providing specific instructions for child dosing.

Finding 4-3: Predispensing provides no flexibility to dispense alternative MCM in case of an attack using a strain of anthrax that is resistant to the predispensed antibiotic. In addition, unlike POD strategies, the development and implementation of a predispensing strategy for the general public would be unlikely to strengthen public health infrastructure and the capability to accomplish other public health goals beyond dispensing antibiotics for anthrax.11

Communications, Behavior, and Adherence

Predispensing MCM to the general public raises many issues related to communications, behavior, and adherence—both outside of an emergency situation and during a public health emergency. Many of these issues apply to all dispensing strategies but may be exacerbated with predispensing.


In the event of a public health emergency, effective communications are critical to ensuring that the exposed and potentially exposed populations receive prophylactic antibiotics. This will be one of the major challenges of the response, regardless of the particular dispensing strategies used. Potential communications-related advantages of predispensing include a sense of individual preparedness that could promote postevent calm, although this has not been corroborated by population-level evidence. In addition, this strategy provides increased time before an event to communicate with and educate the population on using the MCM safely and appropriately. Predispensing also presents specific communications chal1enges for several reasons. First, in addition to telling people when to take their antibiotics, communications would be needed to tell people when not to take their antibiotics, such as in the case of a local public health emergency for which antibiotics were not indicated or an anthrax attack elsewhere in the country. The response to the nuclear accident in Japan highlights that even consistent and pervasive messaging may not be sufficient to achieve this goal. Second, if the anthrax turned out to be resistant to the antibiotic(s) in the home, this would need to be communicated, along with instructions on what to do instead—for example, instructions to go to a POD to receive a different antibiotic. If people felt that they were protected because they had taken their home antibiotics and therefore did not seek additional care, mortality could increase greatly. Finally, MedKits would likely contain only a 10-day supply, so communications would still be needed about how to obtain the remaining doses.

Instructions, health literacy, and limited English proficiency

Comprehensible instructions would be necessary to explain how to take predispensed antibiotics for both children and adults who cannot swallow pills. The study by Stergachis and colleagues (2007) discussed in Chapter 3 raised concern about improper medication use due to a misunderstanding of the instructions provided during rapid dispensing in a nonmedical POD exercise. Similar misunderstandings would likely arise with home prepositioning.

In theory, MedKits may be better suited than personal stockpiles to providing clear, emergency-specific instructions. This advantage is dependent, however, on the FDA's authorizing or approving such instructions. The drug fact sheets included in the EUA MedKits distributed in the Minneapolis-St. Paul postal model pilot are lengthy and written at a high reading level, and translations to other languages are not available (Griffith, 2011). The content of these fact sheets is specified by the EUA and may not be altered for readability or comprehensibility. Similarly, drug product labels are FDA-approved materials that may not be altered. Using MedKits or personal stockpiling effectively and safely as a strategy for a cross-section of the population would require instructions that could be understood by people with low health literacy, as well as translated instructions for non-English speakers.

Social behavior

Little behavioral research has identified the factors that affect individual motivation to participate in preparedness (e.g., perceived risk of attack, cost, at-risk family members); to acquire a home stockpile (e.g., ability to store the MCM); to understand and follow MedKit labeling and instructions; and to self-administer medication from emergency caches without explicit direction that there is a public health emergency (IDSA, 2008; NBSB, 2008). As noted in Chapter 3, communications would have to be developed to effectively reach a sociodemographically diverse audience and would have to involve multiple communications strategies, including both traditional media outlets and social media.


Adherence to the recommended course of prophylactic antibiotics following an anthrax attack is a major concern, regardless of the dispensing method used, as discussed in Chapter 3. No data are available with which to determine whether adherence would be better or worse with predispensed MCM than with PODs. It is conceivable, however, that adherence could be worse for predispensed MCM since there would be no POD staff member specifically emphasizing the importance of taking the full course of antibiotics. In addition, the POD system prevents people from taking MCM during an emergency in which antibiotics are not indicated (since the PODs would not be set up).


Predispensing of MCM has not been tested or used at a scale that would provide the experience needed to fully understand its impact on state and local MCM dispensing plans. A strategy of predispensing MCM for the general public could be government sponsored or could rely on individuals to obtain and own the MCM. The form of the predispensed MCM (e.g., personal stockpile or MedKit) would not greatly impact the logistics of this strategy.

For a government-sponsored predispensing strategy, state and/or local jurisdictions would likely be responsible for the initial distribution, registry, and tracking of the MCM. Jurisdictions would be responsible for ongoing communication with households participating in the plan. Expiring MedKits or personal stockpiles would need to be replaced regularly, imposing another logistic and cost burden on state and local health agencies. Since many states require a 1-year expiration date on prescriptions (regardless of the date indicated by the manufacturer), replacement could be required every year unless a special dispensation could be obtained to allow replacement every 2 or 3 years. A registry would have to be used for notifying households that they should or should not take the MCM; this registry also would need to be maintained.

Privately obtained and owned personal stockpiles or MedKits would shift the responsibility for maintenance to the head of the household. State and local health agencies would not know how many privately obtained predispensed MCM were in the community. Households possessing the kits would not need to use the POD system initially, which would likely alleviate the initial stress on that system. Nevertheless, local jurisdictions still would have to plan for the entire community because (1) they would not know how many households had predispensed MCM; (2) PODs would be needed to provide antibiotics for days 11-60; and (3) if the anthrax attack involved a strain resistant to the predispensed antibiotic, PODs would be needed to provide an alternative antibiotic.

Legal Issues

The legal issues associated with predispensing depend entirely on the form of predispensing used. Personal stockpiles currently are allowed under normal prescribing practices. An EUA MedKit was authorized for dispensing to postal workers participating in the Minneapolis-St. Paul postal model pilot. However, the use of this MedKit is highly restricted. As noted earlier, for example, the first responders accompanying the postal workers did not receive MedKits (FDA, 2009). While HHS, DHS, and others have been discussing a potential EUA covering MedKits for first responders and emergency care providers, this idea has not yet been implemented (IOM, 2010b; NBSB, 2008). A central impediment is that while the USPS is a federal agency and the EUA can be tailored specifically to postal workers, the first responder community varies widely across and within federal, state, and local jurisdictions and has no centralized organizational structure. Given these legal issues, it is unlikely that an EUA MedKit could be implemented for the general public, except, perhaps, for the general public in a limited geographic area if there were intelligence information about a specific, imminent threat.

Currently, no MedKit has been approved by the FDA. The discussion on predispensing below includes the steps that may be required to obtain FDA approval for such a kit.

Predispensing to Targeted Subpopulations

An alternative to using a predispensing strategy for the general public within a community is for predispensing to target specific groups and individuals who would lack access to prophylactic antibiotics via other timely dispensing mechanisms. For these groups and individuals, the risk of not obtaining prophylactic antibiotics following an anthrax attack may outweigh the potential health risks associated with inappropriate use. In addition, with a more limited, targeted strategy, it may be easier to institute systems to decrease inappropriate use, manage costs, and/or develop an alternative dispensing mechanism in the case of an attack with antibiotic-resistant anthrax.

The use of predispensing for some first responders and critical infrastructure workers and for selected patients is discussed here for illustrative purposes, but the benefits and risks of predispensing for specific groups and individuals will vary across communities. In particular, because the committee recommends this strategy primarily for those who lack other means of obtaining MCM in a timely manner, the benefits of this use of predispensing in a community will depend heavily on the coverage of the other dispensing mechanisms in place. State and local public health officials likely already know about groups and individuals that are likely to face challenges in accessing MCM, and they could work with health care providers, local community organizations (e.g., faith-based), and constituent advocacy groups to identify other vulnerable subpopulations and evaluate their access to current dispensing mechanisms.

First Responders and Critical Infrastructure Workers

Certain first responders and critical infrastructure workers (including health care personnel) will be expected to report to work as soon as an attack is detected and/or remain at work for extended periods during the response. Therefore, they will be unable to stand in line at PODs. The HHS et al. (2011) proposal for implementing the postal model includes providing prophylaxis to postal workers' families as well as the workers themselves based on the assumption that workers may not report to work until their family members have been taken care of. While workers may consider other factors in deciding whether to participate in the response, ensuring that these workers have a means of obtaining MCM in a timely manner both yields a multiplicative effect, impacting others in society through the services they provide, and ensures that they have equitable access while fulfilling responsibilities that prevent them from waiting in line at open PODs. Communities can determine how to define these subpopulations and how best to dispense to them.

Most communities will likely find that workplace caches, where feasible, are a better strategy than predispensing for these workers. Workplace caches avoid the risks of inappropriate use described above, provide a better (but not seamless) infrastructure for distributing an alternative MCM if needed, and are more cost-effective (see Chapter 5; note also that in the postal pilot program, MCM were kept in both homes and workplaces, thus increasing costs). However, communities may find that in some cases, workplace caches would not provide timely access while still enabling workers to begin working immediately and to stay at work. These cases might include, for example, first responders and critical infrastructure workers who do not muster in a workplace or workers for whom it would not be feasible to bring MCM from the workplace cache to their family members in a timely fashion. This section outlines considerations for predispensing to these selected groups.

Inappropriate use

There is no available evidence to suggest that first responders and critical infrastructure workers are inherently less likely to misuse antibiotics than members of the general public. However, the results for the St. Louis first-responder cohort and the Minneapolis-St. Paul pilot program suggest that these subpopulations would be unlikely to misuse the MCM during nonemergency times given circumstances similar to those used in the study and pilot program. Of the 1,556 first-responder households in the St. Louis study, 98.7 percent returned their MedKit; 1.3 percent could not locate it (two had thrown it away, and seven had lost it); and one refused to return it (CDC, 2008b). No household in the first responder cohort used the medication. And as noted earlier, 1 year into the postal pilot program, 367 of the 377 kits (97.3 percent) held by active postal volunteers were returned, unopened, for scheduled replacement (Griffith, 2011). Unlike predispensing programs for the general public, programs set up through employers may provide more screening, supervision, and incentives not to misuse the medication, thus being more aligned with the St. Louis study and postal pilot program.

Although the data suggest that these groups may not misuse the MCM in nonemergency situations, no data are available on whether these groups would misuse them in an emergency situation in which antibiotics were not indicated—for example, a nonanthrax public health emergency, a false alarm caused by a nonanthrax white-powder event, or an anthrax attack far away. As with the general public, there also is no appropriate evidence with which to address the question of whether the form of predispensing would impact the rate of inappropriate use among first responders.


Predispensing for some first responders and critical infrastructure workers raises concerns about flexibility similar to those described above for the general public, particularly with regard to what would happen if the anthrax used in an attack were resistant to the predispensed MCM. However, this issue may be somewhat mitigated since it may be possible to develop a plan through which employers would communicate with employees about the issue (e.g., through a phone tree) and distribute an alternative MCM in the workplace.

Legal considerations

Jurisdictions, in partnership with public and private employers as applicable, could implement a home-based predispensing strategy using personal stockpiling, which currently would be allowed under normal prescribing laws. Such a strategy also could be implemented with an FDA-approved MedKit if one were to be developed and licensed. As noted above, however, there is limited evidence to suggest that the use of such a kit would decrease the rate of inappropriate use in these subpopulations; therefore, it is not clear whether the considerable expense of developing an FDA-approved MedKit would be worthwhile. The costs associated with developing an approved MedKit are touched on below.

As an intermediate solution, it might be possible to use a MedKit authorized under an EUA (as was done in Minneapolis-St. Paul), although accomplishing this may not be straightforward. During a discussion at a 2009 IOM workshop, it was noted that although an EUA could be obtained for postal workers who volunteered to participate in the postal pilot, obtaining similar consideration for the first responders who would accompany the postal workers as they delivered MCM postattack was challenging (and has to date not been accomplished) (IOM, 2010b). In particular, speakers observed that all postal workers have a single employer (the USPS), whereas first responders are employed by jurisdictions at many different levels and privately. Adding critical infrastructure workers to the mix would only increase the challenges.

Selected Patients

Certain patients may have social situations and/or medical conditions that preclude them (or are a significant barrier) from accessing medications through the public health system. For example, some patients might be unable to travel to PODs or might have a compromised immune system that would make it unadvisable to stand in line with a crowd. In many cases, such patients could rely on another household adult or a neighbor to obtain MCM, and some jurisdictions might develop plans through which home health care workers or other service delivery agencies would deliver MCM. For patients without access via these mechanisms, the potential risk of not having antibiotics following an anthrax attack may outweigh concerns about health risks from inappropriate use, lack of flexibility, and cost.

Kent Sepkowitz, Vice Chairman of Clinical Affairs at Memorial Sloan-Kettering Cancer Center, suggested to the committee that oncology patients and possibly HIV patients might appropriately receive predispensed antibiotics. These patients would not be well suited to standing in POD lines, and they have a long history with prepositioned antibiotics and their use with appropriate physician control (Sepkowitz, 2011). Little available published evidence exists on whether patients with complex medical conditions are less likely to misuse antibiotics relative to the general population, and there is likely to be a great deal of individual variation in this regard. Physicians considering prescribing antibiotics for protection against anthrax would need to take into account individual patients and their demonstrated level of adherence to medication instructions. For such patients, predispensing also would make it possible to adapt to their individual needs. For example, alternatives to doxycycline could be provided for those allergic to that drug, and the time would be available to provide thoughtful solutions to those at risk of drug interactions or with complex physiology.

Predispensing for these selected patients is most likely to entail personal stockpiling through normal physician-patient contact and prescription routes, as an independent activity. Thus it would be done on an ad hoc basis and would likely be covered by health insurers.

The committee is not recommending predispensing for those who have other ways of obtaining access to postexposure prophylactic antibiotics given the concerns discussed above about potential health risks, lack of flexibility, and cost. Predispensing is warranted only if the alternative (i.e., no access to antibiotics postevent) is worse. Specifically, the committee is not recommending predispensing for those who are anxious about an anthrax attack but who could, for example, obtain MCM at a POD, although the committee recognizes that this is currently allowed legally under normal prescription laws. The committee acknowledges that public health officials have not done a very good job at communicating what plans are in place, and therefore it may be challenging for physicians and their patients to determine whether and how a patient would access antibiotics in case of an anthrax attack. Improving communication on existing dispensing plans may be a safer and more effective means of decreasing anxiety than promoting widespread availability of predispensed MCM. Table 4-3 provides a summary and comparison of the different potential forms of predispensing of MCM: the general safety-related advantages, the potential for inappropriate use among the general population and target subpopulations, and associated costs.

TABLE 4-3. Comparison of Potential for Inappropriate Use Among Different Potential Forms of Predispensed Antibiotics.


Comparison of Potential for Inappropriate Use Among Different Potential Forms of Predispensed Antibiotics.




The term cache often is used broadly to describe stockpiles of MCM held by state or local jurisdictions, health care facilities, and private-sector organizations, among others. For the purposes of this report, and to enable clear discussion of the different properties associated with different types of prepositioning, the committee defines cache more specifically to denote storage of MCM in the locations from which they will be dispensed and uses the term stockpile to cover federal, state, and local stockpiles.


Since the passage of the Pandemic and All-Hazards Preparedness Act in 2006, the HPP has been administered by the Office of the Assistant Secretary for Preparedness and Response (ASPR). The program was originally administered by the Health Resources and Services Administration (HRSA), during which time it was called the National Bioterrorism Hospital Preparedness Program (ASPR, 2011a).


Personal communication, John Hick, medical director for emergency preparedness, Hennepin County Medical Center, MN, June 15, 2011.


Section 104 of the Model Pharmacy Act, a version of which has been adopted by most states, defines the practice of pharmacy to include “proper and safe storage of Drugs and Devices” (among other things). Private organizations caching antibiotics on-site would fall within this definition in the absence of an emergency declaration and issued EUA, and therefore be required to obtain a permit/license to “practice” pharmacy.


Federal Law: Food, Drug, and Cosmetic Act (21 U.S.C.). Rockville, MD: FDA, http://www​.fda.gov/RegulatoryInformation​/Legislation/FederalFoodDrugandCosmeticActFDCActucm086299​.htm#chapV (accessed August 5, 2011). State Law: NABPLAW® database search of state “prescription labeling requirements,” conducted in November 2008 and provided to the committee by E. Lewalski, National Association of Boards of Pharmacy, July 18, 2011.


Personal communication, J. Hodge, Public Health Law Network, June 22, 2011.


Personal communication, Linda Neff, lead study author and senior epidemiologist, CDC, June 17, 2011.


Along with other findings presented in Chapter 5, Finding 4-2 leads to Recommendations 5-4 and 5-5 in Chapter 5.


FDA-approved labeling for doxycycline outlines the following adverse reactions: “Due to oral doxycycline's virtually complete absorption, side effects of the lower bowel, particularly diarrhea, have been infrequent. The following adverse reactions have been observed in patients receiving tetracyclines: Gastrointestinal: anorexia, nausea, vomiting, diarrhea, glossitis, dysphagia, enterocolitis, and inflammatory lesions (with monilial overgrowth) in the anogenital region. Hepatotoxicity has been reported rarely. These reactions have been caused by both the oral and parenteral administration of tetracyclines. Rare instances of esophagitis and esophageal ulcerations have been reported in patients receiving capsule and tablet forms of the drugs in the tetracycline class. Most of these patients took medications immediately before going to bed. … Skin: maculopapular and erythematous rashes. Exfoliative dermatitis has been reported but is uncommon. Photosensitivity is discussed above. … Renal toxicity: Rise in BUN has been reported and is apparently dose related. … Hypersensitivity reactions: urticaria, angioneurotic edema, anaphylaxis, anaphylactoid purpura, serum sickness, pericarditis, and exacerbation of systemic lupus erythematosus. Blood: Hemolytic anemia, thrombocytopenia, neutropenia, and eosinophilia have been reported. Other: bulging fontanels in infants and intracranial hypertension in adults. … When given over prolonged periods, tetracyclines have been reported to produce brown-black microscopic discoloration of the thyroid gland. No abnormalities of thyroid function studies are known to occur” (Pfizer, 2011, pp. 12–13).


Along with other findings discussed in Chapter 5, Finding 4-3 leads to Recommendations 5-4 and 5-5 in Chapter 5.

Copyright 2012 by the National Academy of Sciences. All rights reserved.
Bookshelf ID: NBK190049


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (3.7M)

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...