Development of the Original Framework

Substantive and Procedural Values

The working group also proposed a set of fundamental ethical values to consider when making decisions about allocating scarce federal public health and medical resources. The work group based the framework on the literature on medical and public health ethics and theories of distributive justice and applied them to decisions regarding scarce resource allocation. The intent was to reflect the values of society and existing federal guidance.

Substantive values are strongly held ethical ideals that inform decisions and actions (ie, decisions and actions about what is right or should be done in the face of uncertainty or conflict). ⁸ The substantive values the work group identified are shown in Table 1. Because consensus on substantive ethical principles is often difficult to achieve, it is generally recognized that scarce resource allocation decisions must also be based on fair procedures. Such procedures help ensure that even when agreement is not unanimous regarding the allocation decisions themselves, stakeholders will recognize that the decisions resulted from processes that are open, reasonable, inclusive, and fair. Procedural values, therefore, are ethical ideals that promote and can be used to evaluate the fairness of a process for decision making under uncertainty or conflict. Ensuring a procedurally fair process is important precisely because stakeholders may differ about substantive values. The procedural values that the work group identified are shown in Table 2.

Table 1 Substantive Values to Guide Allocation Decisions in Disaster Situations ^{5 , 8 ,} Reference Persad, Wertheimer and Emanuel ^{9 ,} Reference Vawter, Garrett and Gervais ^{11 , 13 -} Reference Roberts and DeRenzo ¹⁵

Table 2 Procedural Values to Guide Allocation Decisions in Disaster Situations ^{5 , 8 ,} Reference Persad, Wertheimer and Emanuel ^{9 ,} Reference Vawter, Garrett and Gervais ^{11 , 13 -} Reference Roberts and DeRenzo ¹⁵

Concepts Consistent With the Principle of Federalism

Along with these substantive and procedural values, guidelines for fair allocation of federal resources should take into account political theories about federalism—the division of sovereign authority among levels of government. ²¹ As described in Table 3, federal leaders should apportion resources to disaster-affected states in a manner that takes into account features of a well-functioning federal system. ²²

Table 3 Concepts Consistent with the Principle of Federalism ^{1 , 2 ,} Reference Buchanan ^{23 , 24}

Abbreviations: ASPR, Assistant Secretary for Preparedness and Response; CDC, Centers for Diseases Control and Prevention.

To support states, federal leaders should recognize the role and perspective of state leaders, as they are likely to be best informed about their situations. At the same time, federal leaders are likely to be able to compare the relative needs when several states make requests for assistance. Given this complex reality, both state and federal governments have valuable perspectives to offer regarding the amount of resources that each state needs to respond to a disaster and the amount that the federal government should offer in assistance. Accordingly, the decision about the amount of scarce resources to be given to each state should consider that states with the greatest needs may experience a reduced ability to govern and communicate and, therefore, may submit requests after states with lesser needs.

Recognizing the important partnership between federal and state leaders in addressing disasters most effectively, the National Response Framework specifies that operational planning needs to integrate federal departments and agencies and other national-level partners to provide the right resources at the right time to support state response operations. ²

Allocation Criteria and Scoring System

The substantive and procedural values and concepts consistent with the principle of federalism provided the basis for 15 allocation criteria that the work group proposed federal decision makers might use to guide decisions regarding allocation of scarce federal public health and medical resources. The criteria and their definitions are summarized in Table 4. Initially, the work group assigned equal weight to each of the criteria and applied a scoring system to quantify the criteria and to provide a rigorous approach to the allocation decisions. With use of the scoring system, each affected state would be evaluated according to the 15 allocation criteria on a scale of 0 to 2, to produce a total score for each state. Prioritizing among states would be based on the total score each state received when assessed against all 15 criteria. States would not be compared with one another according to any individual criterion. The rating specifications and scoring are also shown in Table 4.

Table 4 Criteria, Rating Specifications and ScoringReference Danis, Hansen and Knebel ^{27 -} Reference Sharpe, Danis and Knebel ²⁹

^aFor all population-based criteria (1 to 7), decision makers should use a consistent approach (either state or county data) in scoring.

^bN = number of people in state/county.

Stakeholder Engagement Workshops

We presented the criteria and the scoring system to stakeholders to test the validity of the proposed approach to federal resource allocation decision making. The engagements included the following groups: (1) disaster preparedness and emergency management experts attending the annual Integrated Medical, Public Health, Preparedness and Response Training Summit (ITS)Reference Danis, Hansen and Knebel ²⁷ ; bioethicists attending the annual meeting of the American Society for Bioethics and Humanities (ASBH)Reference Sharpe, Danis and Knebel ²⁹ ; and attendees of the National Institutes of Health (NIH) Bioethics Interest Group.Reference Danis, Sharpe and Knebel ²⁸

Approximately 130 disaster response experts participating in the ITSReference Danis, Hansen and Knebel ²⁷ chose to attend a breakout session in which the allocation framework was discussed. Participants were asked to rate, using audience response technology, each of the allocation criteria on a 5-point Likert scale ranging from “extremely important,” to “not important, can be eliminated.” Six criteria (1,3,8,10-12) had the highest percentage of people who ranked them as extremely important, very important, or important (Table 4). The attendees also rated 2 of the criteria (9 and 15) as unimportant and could be eliminated (Table 4). The remaining criteria received ratings of moderate importance. These findings support the face validity of most of the criteria among disaster response experts. Open-ended comments supported the finding that the respondents thought that some criteria were more important than others, suggesting that the criteria should not be weighted equally.

The engagements with the ASBH and the NIH Bioethics Interest Group used a discussion format, rather than audience-response technology. During the workshops, we provided the background of the problem and the proposed framework and a hypothetical disaster scenario to focus the discussion of the criteria and asked participants to deliberate about the following:

• • the usability and acceptability of the allocation criteria

• • whether any other criteria should be added

• • the utility of the proposed scoring system.

Participants at both of these meetings commented that the framework provided ethically useful guideposts for allocating scarce federal resources and that the criteria reflected the procedural and substantive values relevant to ethical disaster planning and response. However, many respondents also found the criteria to be vague or cumbersome to apply in the context of the hypothetical disaster scenario. They also found the criteria difficult to apply when the data needed to score them were not available. Some agreed with the stakeholders from the ITS and suggested that the criteria should not be weighted equally.

Logic Model

Following these stakeholder engagements, we consulted the analytic decision support group of the Biomedical Advanced Research and Development Authority (BARDA) regarding next steps in the development of the framework, given their expertise in modeling allocation decision making. The BARDA group proposed a logic model to illustrate the decision-making process using the criteria proposed in the original framework (Figure). One difference, however, is that while the original criteria were focused on specific resources, the logic model can be applied to specific resources or to resources more generally. The logic model combines a number of the original criteria and focuses on a qualitative rather than a quantitative approach.

Figure Logic Model Original work developed in collaboration with modeling experts from the Analytic Decision Support Group, Biomedical Advanced Research and Development Authority, US Department of Health and Human Services, Office of the Assistant Secretary for Preparedness and Response (Appendix).

Moreover, the logic model retains the criteria identified in the original framework, which stakeholders agreed were ethically relevant, but organizes the criteria in the form of a decision procedure that may be easier to employ. In this approach, the logic model addresses the concern raised by stakeholders that quantitative scoring of the criteria was cumbersome and that it was difficult to use when situational awareness is limited. It also addresses concerns raised by the BARDA group that a quantitative approach implies a level of precision that is not possible in this type of decision making. The logic model's qualitative approach does not attribute equal weight to the criteria; instead, it allows decision makers to weight relevant criteria differently, as appropriate to the circumstances of the disaster.

The logic model starts with the presumption of equal allocation, which is a modification of criterion 15 from the original framework that specified equal consideration for each state. The presumption is that all affected states are entitled to the same allocation of the scarce federal resources under the same circumstances. This provisional allocation may indicate that each state gets the same amount of the resource or that each state's request is reduced by the same amount. If sufficient differences in need occur, rendering an equal allocation of the scarce resource unfair, then the provisional allocation should be modified based on the considerations in subsequent steps of the logic model. Federal decision makers would continue to the next steps in the logic model for needed refinements to equal allocation.

At the second step of the logic model, decision makers modify the provisional allocation according to the population at risk. This step combines criteria 1 through 4 from the original framework. While the scarce resource need not be allocated solely based on the number of people affected by the disaster, the need for the resource and the potential for the resource to benefit people are both highly correlated with the number of people affected. The types and severity of injuries and the potential for follow-on injuries are both legitimate considerations for allocating the resources. Different types of scarce resources do not need to be equally distributed if it would produce an unfair or inefficient allocation. For example, if a state has a high number of burn injuries while another has a large number of crush injuries, assigning each state an equal number of burn and crush injury teams would not be helpful or ethical. If no meaningful difference in the population at risk is found in the affected states, no adjustment to the provisional allocation is needed at this step.

At the third step of the logic model, decision makers will make modifications to account for destruction of infrastructure (including critical infrastructure and key resources). This step combines criteria 8 and 10 from the original framework and adds critical infrastructure and key resources that have a local impact. For example, if a nuclear power plant is damaged during an earthquake, local impact will occur from radiation exposure; also, national implications will arise because of decreased energy production. If no meaningful difference is found in the degree of destruction of infrastructure between states, then no adjustment to the provisional allocation is needed at this step.

The fourth step of the logic model considers vulnerable populations. In the original framework, criteria 5 through 7 considered special needs populations. With the logic model the language was changed to vulnerable populations to align terminology with existing guidance. ³² If the difference in vulnerable populations is significant between different states, the allocation of the scarce resources may need to be adjusted to make the allocation more equitable. Potential vulnerable populations may include, but are not limited to, the following:

• • Individuals with pre-existing access needs or functional, medical, mental health, or psychological needs for care, treatment, or pharmaceuticals, ³²

• • Segments of the population below the poverty threshold, and

• • Medically underserved populations.

If no meaningful difference in vulnerable populations exists among the states, no adjustment is needed to the provisional allocation at this step.

The fifth step in the logic model considers whether the scarce resource is likely to benefit those affected. This step combines criteria 9 and 11 from the original framework. Because the purpose of resource allocation is to help those affected by the disaster, the resources that are provided must be pertinent to the local conditions. If local conditions prevent a scarce resource from being effectively delivered or used due to destruction of required infrastructure, damage to transportation networks that negates the ability of the aid to arrive in a timely manner, or other conditions that render the aid unhelpful, then the provisional allocation must be adjusted so the scarce resource can be sent and used effectively elsewhere. As an example, medical teams that are meant to augment hospitals may not be used effectively if the hospitals are severely damaged. In that circumstance, the teams may be better deployed to a state that can use this scarce resource more effectively. A separate determination would need to be made about whether self-sufficient medical teams are available for deployment to the state that suffered destruction of its hospitals. In making this assessment, federal leaders acknowledge the perceptions and perspectives of state leaders regarding their needs by taking into consideration state requests.

The sixth step of the logic model accounts for the availability of other sources of aid, combining criteria 12 and 14 from the original framework. Federal agencies do not provide aid in a vacuum. When a particular federal resource or resources more generally are scarce, the allocation decision should take into account other aid being made available to affected states. If a state has an internal or other non-federal source for the needed capability, the state's need for the federal resource is decreased. In this case, it would be fairer to decrease the amount of the resource provided to that state and increase the allocation to states that do not have an external source of aid.

The last step of the logic model addresses other considerations (including critical national priorities), which is criterion number 13 in the original framework. These considerations may include White House policies, statutes, regulations, and critical infrastructures and key resources. Examples include presidential direction with regard to international evacuation; legislation regarding spending of disaster relief dollars; US Food and Drug Administration regulations with regard to international donations of food and medications; and damage to critical infrastructures that do not have a local public health and medical impact but could have national implications such as a data center that supports financial services. When analyzing their decisions, federal decision makers should consider factors that have not yet been addressed. If the allocation resulting from the 6 previous steps appears unfair, given such considerations, it should be revised.