Clinical Care in Disasters

Disaster events will be marked by a sudden or gradual increase in demand for healthcare services and a related decrease in the supply of resources available to provide such care. This will result in a healthcare-sector response that requires implementation of a variety of “surge capacity” strategies that include steps taken to reduce demand for care (e.g., the implementation of community-based triage capabilities and risk communication about when to seek care) and the augmentation of ambulatory care capacity in addition to better described inpatient care strategies (Hick et al., 2004; Kaji et al., 2006; Barbisch and Koenig, 2006; Davis et al., 2005; Kelen et al., 2006, 2009; California Department of Public Health, 2008 ; Hanfling, 2006). Therefore, all healthcare entities—not just hospitals—should have plans to provide crisis care. Outpa tient facilities (and community-based clinics, nursing homes, primary care, etc.) may use strategies modified from hospital guidance. EMS agencies may elect to use sample strategies as outlined below or develop system-specific responses.

A number of strategies can be used to bolster the supply of key resources (i.e., space to deliver care, clinical staffing availability, and the availability of key supplies) (Hick et al., 2009; Kaji et al., 2006; Hick et al., 2009). Most likely the crisis will occur over a spectrum of supply and demand spikes, suggesting that a continuum of care will be in place over the course of any disaster response. It may be helpful to consider that surge capacity following a mass casualty incident falls into three basic categories, depending on the magnitude of the event: conventional, contingency, and crisis surge capacity (Box 2). Note that the same event may result in conventional care at a major trauma center, but crisis care at a smaller, rural facility.

BOX 2

Conventional, Contingency, and Crisis Capacity. Conventional capacity—The spaces, staff, and supplies used are consistent with daily practices within the institution. These spaces and practices are used during a major mass casualty incident that (more...)

Conventional, contingency, and crisis care represent a continuum of patient care delivered during a disaster event. As the imbalance increases between resource availability and demand, health care—emblematic of the healthcare system as a whole—maximizes conventional capacity, then moves into contingency, and, once maximized, moves finally into crisis capacity. Concurrent with this transition along a surge capacity continuum is the realization that the standard of care will shift. This occurs primarily as a result of the growing scarcity of human and material resources needed to treat, transport, and provide patient care. The goal of the healthcare agency or facility is to return as quickly as possible to conventional care by requesting resources or transferring patients out of the area, drawing on the resources of partner or coalition hospitals and the health system as a whole. Along the span from conventional to crisis care, healthcare facilities should attempt to minimize changes that significantly impact patient outcomes by changing work practices in order to focus resources on patient care (Phillips and Knebel, 2007; ANA, 2008; Gebbie et al., 2009) (Figure 1).

FIGURE 1

Continuum of incident care and implications for standards of care. NOTE: Post anesthesia care unit (PACU); intensive care unity (ICU)

Catastrophic events will have an impact on the entire healthcare delivery “system” and will affect response and delivery of care that occurs in the home, community, hospitals, primary care offices, and long-term care facilities. A number of strategies can be implemented along this continuum of care delivery to reduce the likelihood that standards of care will change in a disaster situation. These include steps taken to substitute, conserve, adapt, and reuse critical resources, including the way staff are used in delivering care. All these steps should be attempted prior to the reallocation of critical resources in short supply (Tables 3 and 4). Every attempt must be made to maintain usual practices and the expected standard of care and patient safety (Rubinson et al., 2008; Minnesota Department of Health, 2008).

TABLE 3

Sample Strategies to Address Resource Shortages.

TABLE 4

Sample Strategies for Emergency Medical Services (EMS) Agencies to Address Resource Shortages.

Broadening surge capacity must incorporate the full spectrum of patient care delivery capabilities in a disaster-impacted community. This includes planning for extension of hospital-like services in an unregulated, non-healthcare setting. Examples of this include the establishment of Federal Medical Stations (FMSs) during the responses to the multiple Florida hurricanes in summer 2004, Hurricanes Katrina and Rita in 2005, and Hurricanes Gustav and Ike in 2008 (HHS, 2009). The initial concepts for such planning came from work conducted for the U.S. Army Soldier Biological Chemical Command in the late 1990s. These efforts focused on a combination of out-of-hospital capabilities divided between Neighborhood Emergency Help Centers (NEHCs) and Acute Care Centers (ACCs) (Church, 2001a, 2001b; Skidmore et al., May 2003; AHRQ, December 2004; Hamilton et al., 2009a; Hamilton et al., 2009b; Gavagan et al., 2006).

The NEHC is intended to function as a community care station that provides a combination of functions, including victim triage, and serves as a distribution point for medical countermeasures. The ACC, similar to the FMS concept, serves as an out-of-hospital medical treatment facility for patients requiring a lower acuity level of care than that supported in a hospital critical care setting, but not well enough to be managed at home. Pandemic influenza planning has galvanized many communities to adopt such an approach to surge capacity planning, largely based on this theoretical framework (Cinti et al., 2008). The components of this alternate care system are built around a stratification of care model, with emphasis on the use of triage algorithms that prioritize use of community-based services for selective patient care delivery that might otherwise be managed under non-disaster circumstances in the hospital setting. The committee has made the assumption that the delivery of care in an unregulated environment would be construed as an alteration to the existing standard of care. Yet such an approach may be necessary in order to prevent collapse of overburdened hospitals responding to a surge event. Even absent the threat of collapse, in some circumstances (such as an infectious epidemic) it is possible that higher quality, safer care can be provided outside the usual venues for most patients. In such conditions, a decision to relocate most care from hospital emergency departments to alternate care facilities would comprise a change in the usual standard of care, but superior quality compared to attempting to maintain ordinary use of the usual facilities.

Disaster Mental Health Crisis Standards of Care

In major disaster and emergencies, there will also be a surge of psychological casualties among those directly affected, including responders, healthcare practitioners, and members of the population who have not experienced direct impact. Mass psychological casualties and morbidity will occur in those who experience an aggravation of a prior or concurrent mental health condition. New substantial burdens of clinical disorders, including PTSD, depression, and substance abuse may also arise among those with no prior history. Even in those with no formal disorder, there may be significant distress at a population level, resulting in unparalleled demands on the mental health system.

The magnitude of new incidence disorder in the population has typically ranged from 30 to 40 percent or more in those directly impacted, such as those who experienced personal losses (IOM, 2003; Galea et al., 2005). Although resilience may also be a result for some, the population-level impact of mass casualty incidents compared to other types of disas ter will likely result in an substantive mental health burden on the nation during and after use of crisis standards of care requiring mental health interventions across varied “disaster systems of care” including the healthcare system, public and private mental health systems, schools, and coroner and other key systems at the community level (Schreiber, 2005).

Therefore, it is necessary to use a mass casualty disaster mental health concept of operations to enable a crisis standard of disaster mental health care through the use of currently available evidence-based mental health rapid triage and incident management systems. For example, such systems used by the American Red Cross and Los Angeles (LA) County Emergency Medical Services Agency and those recommended by the National Biodefense Science Board Disaster Mental Health workgroup may serve as models (HHS, November 2008). The latter system, known as “PsySTART,” provides for rational allocation and alignment of limited acute- and response-phase mental health assets to those with greatest evidence-based risks and needs in a phased, sequential manner so that those in need are matched to resources in the most timely fashion during response and recovery (Thienkrua et al., 2006). In the Los Angeles County Emergency Medical Services agency pilot project, for example, Los Angeles’s network of 14 Disaster Resource Center Hospitals (Level 1 trauma centers), the Department of Mental Health, and other key “disaster systems of care” collect and are able to share triage information for near real-time situational awareness and a “common operating picture.” This information guides prioritization of crisis intervention at the hospitals and facilitates mutual aid across NIMS levels. Those with the greatest triaged needs are matched to available care until all those who are at risk and desire services can be further assessed and linked in the most timely manner to definitive care (Schreiber, 2005). There is now evidence that certain types of psychological interventions are the treatments of choice for conditions such as PTSD that are a frequent result from disasters, and the triage system allows for faster matching of the high-risk subset to appropriate and timely care (IOM, 2007).

Palliative Care Planning for Crisis Standards of Care

The provision of palliative care in the context of a disaster with scarce resources is a relatively new component of disaster planning. The goal of palliative care is to prevent and ease suffering and to offer patients and their families the best possible quality of life at any stage of a serious or life-threatening illness and is not dependent on prognosis. It can also be provided at the same time as curative and life-prolonging treatment.

Although the primary goal of a coordinated response to a disaster incident should be to maximize the numbers of lives saved, a practical plan also must provide the greatest comfort for those who will live for awhile before dying as a result of the incident (Holt, 2008). Triage and treatment practices that focus on maximizing the number of lives saved means that during a crisis, some people who might be successfully treated or cured under normal circumstances will die. During a crisis, palliative care would provide aggressive treatment of symptoms, such as pain and shortness of breath. In addition, triage to palliative care should allow for the fact that the initial prognosis for some patients will change, whether by virtue of their doing better than expected or by additional treatment resources becoming available.

Identifying transition points in the condition of patients helps the patient, family, and healthcare providers prepare for the final stage of life. A transition point can be defined as an event in the trajectory of an illness that moves the patient closer to death. For example, a patient with chronic obstructive pulmonary disease may have no change in her condition until she gets influenza and never fully recovers; for that patient, contracting influenza is a transition point in her condition (Berry and Matzo, 2004). Prognostication, aided by a risk index or scale, enables healthcare practitioners to plan clinical strategies during a crisis situation. These tools may be helpful in determining whether a patient’s illness has reached a terminal phase (Box 3) (Matzo, 2004).

BOX 3

Palliative Care Triage Tools. Flacker Mortality Score: Flacker and Kiely developed a model for identifying factors associated with one-year mortality (the probability of death within the next year) by conducting a retrospective cohort study using Minimum (more...)

Providing a treatment category of “palliative care” for those not likely to survive will be an important service option for responders and triage officers. Acknowledging that a patient is not likely to survive typically leads to discussions regarding the goals of care, appropriateness of interventions, and efforts to help the patient and family begin to say good-bye (Matzo, 2004).

When resources are scarce, planners can make available alternative means of palliative care delivery and treatment. Planners should:

• Develop evacuation plans for existing and new palliative care patients;
• Develop a community response plan, staffing plans, and training programs for first responders and other relevant medical personnel;
• Establish transparent, community-based, and explicit triage criteria for casualties not likely to survive;
• Develop a community education program to prepare the public;
• Stockpile needed palliative care medications and supplies (Wilkinson and Matzo, 2006); and
• Participate in disaster planning, response and recovery training, and public education (Holt, 2008).

Crisis Standards of Care Indicators

Resources that are likely to be scarce in a crisis care environment and may justify specific planning and tracking include:

• Ventilators and components
• Oxygen and oxygen delivery devices
• Vascular access devices
• Intensive care unit (ICU) beds
• Healthcare providers, particularly critical care, burn, and surgical/anesthesia staff (nurses and physicians) and respiratory therapists
• Hospitals (due to infrastructure damage or compromise)
• Specialty medications or intravenous fluids (sedatives/analgesics, specific antibiotics, antivirals, etc.)
• Vasopressors/inotropes
• Medical transportation

Implementation of crisis standards of care first requires recognition of a resource shortfall or impending resource shortfall. However, good situational awareness and incident management can often forestall any requirement to adjust standards of care as patients can either be moved to areas with resources or resources can be brought in to ameliorate the shortage prior to significant consequences for the patient(s). The committee recognizes that this is a particularly important issue for rural healthcare facilities. This is facilitated by monitoring critical resources and evolving events (e.g., ICU bed availability, ventilator availability, and other external health system measures such as situational awareness of both illness and injury numbers and rates within the community, epidemic curve modeling, etc.) for indicators of the need for additional resources or, if no resources are available and no adaptive strategies can be implemented, planning for crisis standards of care. If there is a “no-notice” event such as a major explosion, or indicators are not available (or adjustments are not made or not able to be made), trigger events may occur (Box 4).

BOX 4

Indicators and Triggers. Indicator—measurement or predictor that is used to recognize capacity and capability problems within the healthcare system, suggesting that crisis standards of care may become necessary and requiring further analysis or (more...)

Indicators such as bed availability are tracked routinely by many hospital systems, and surveillance tools monitor other data streams to provide possible early clues to an evolving epidemic. In addition to event-specific data tracking (e.g., ventilators), these indicators should be used where available to determine the “cushion” within the healthcare system and its variability over time.

Facility, local, and regional indicators should be developed to enable anticipation and management of an incident prior to resources being overwhelmed. When event information is not available before it occurs, a system should be in place to collect/share that information during an event. Indicators may also be needed in the out-patient, homecare, and other environments, but have not yet been described.

The committee was unable to identify evidence that specific indicators have predictive value for intervention (Schultz and Koenig, 2006; Davidson et al., 2006; McCarthy et al., 2006), thus, the indicators noted in this document represent expert opinion only, and should be the subject of further research. Due to variables in staffing, in-patient census, and system characteristics, there were no data points that qualified as “triggers” for automatic action absent a sudden overwhelming event that would not require indicators to recognize. The members did feel strongly, however, that waiting for hard “trigger” evidence of crisis care was inappropriate, and that the goal should be anticipation of resource shortages based on situational awareness (including tracking of indicators), with correction of the problem prior to crisis when possible. The numbers reflected in the table are examples only, as there is tremendous variability between regions (Table 5). For example, at the workshop hosted by the committee some panelists believed that one hospital on ambulance diversion should be an indicator, while others noted that multiple hospitals were on diversion on a routine basis in their communities.

TABLE 5

Possible Indicators for Crisis Capacity.

There was agreement with the panelist that 18 hospitals on divert during the severe heat wave in Chicago certainly met the qualification of “indicator” (Stein-Spencer, 2009). In addition, staff absenteeism is likely to affect rural facilities and services disproportionately more than larger urban facilities, and “indicator” thresholds for the impact of infrastructure damage also will vary substantially. Despite the lack of specificity available to the committee, we describe opportunities for indicator capture in the hopes that further study may allow better definition of meaningful thresholds that may have at least some applicability across different populations. In particular, the committee acknowledges that triggers to move to crisis standards of care will likely be different for rural versus urban regions of a state. Therefore, this issue needs to be considered when formulating crisis standards of care protocols for use in disaster situations.

Trigger events revolve around changes to staff, space, and supplies that constitute a change in standard practices such that morbidity and mortality risks to the patient increase (i.e., to crisis standards of care). Trigger events do not necessarily require a state response. If the institution rapidly receives victims from a bomb blast that result in temporary (hours) use of cots for stable patients, but is able to return to conventional operations quickly, the facility can manage this incident internally without the need for the declared crisis standards of care. However, most such incidents require engagement of other healthcare facilities to distribute patients to hospitals with more adequate resources. An example is the 2003 Rhode Island nightclub fire, when manual ventilation of patients was performed in hallways pending air evacuation to regional burn centers (Dacey, 2003). Only in the case that the trigger event(s) are unable to be ameliorated by patient evacuation or resource acquisition is state action required to provide protections to providers who are now delivering care under crisis conditions. This may occur in catastrophic events causing significant infrastructure loss and impeding patient trans port (major hurricane or earthquake) or an epidemic (e.g., pandemic) that affects all institutions.

Trigger points are only reached when the institutional surge capacity cannot accommodate the demand through conventional or contingency responses that do not require an adjusted standard of care (Table 6). Trigger points and actions taken when they occur can be easily incorporated into job action sheets or surge capacity templates used at a hospital (e.g., “if providing cot-based care, hospital must notify Regional Medical Coordination Center (RMCC) by calling [555-555-5555]”). Regional personnel monitoring indicators and triggers must also have easy, intuitive, scripted responses to a notification. Some regions may use categorical systems, but these require significant training and maintenance to be effective and understood, and are best used in well-developed, metropolitan systems (University of California, Davis, et al., 2009).

TABLE 6

Possible Triggers for Adjusting Standards of Care.

Crisis Standards of Care Implementation Criteria

Prior to implementation of formal resource triage, the following conditions must be met or in process (Devereaux et al., 2008b):

• Identification of critically limited resources and infrastructure
• Surge capacity fully employed within healthcare facility
• Maximal attempts at conservation, reuse, adaptation, and substitution performed
• Regional, state, and federal resource allocation insufficient to meet demand
• Patient transfer or resource importation not possible or will occur too late to consider bridging therapies
• Request for necessary resources made to local and regional health officials
• Declared state of emergency (or in process)

Crisis Standards of Care Triage

Triage occurs routinely in medicine, when resources are not evenly distributed or temporarily overwhelmed. Examples include transfer of a patient to a trauma or burn center because most hospitals do not specialize in these types of care, or a mass casualty incident when priority must be assigned for diagnostic imaging or surgery. These decisions are generally ad hoc, based on provider expertise, and have minimal effects on patient outcome. Thus standards of care are routinely adjusted to resources available to the provider without requiring a formal process or declarations. However, the situation in disasters is more complex, as services the hospital usually provide may not be available or not available at all due to demand, with severe consequences to the patient who does not receive these resources.

Triage involves both an assessment of the patient’s condition and the available resources. Triage of patients may occur at three points over the course of patient care: (1) primary triage—triage of patients at first contact with the medical system (dispatch, EMS, or emergency department, at which point patients are assigned an acuity level based on the severity of their illness/disease); (2) secondary triage—reevaluation of the patient’s condition after initial medical care (this may occur at the scene of the disaster or at the hospital following EMS interventions or after initial interventions in the emergency department); and (3) tertiary triage—reevaluation of the patients’ response to treatment after further interventions that may continue during their hospital stay. This is the least practiced, least well-defined and perhaps most ethically challenging type of triage since it might entail removing a life-sustaining resource from one patient in order to provide it to another who is more likely to survive. Such decisions will always be wrenching, regardless the degree of one’s training and preparation. Making them in an ad hoc fashion, without careful clinical and ethical consideration and guidance, is extremely risky.

Furthermore, triage is different during two distinct response phases: reactive triage and proactive triage (Table 7). Reactive triage involves the ad hoc decisions made by clinical or administrative personnel to an exigent circumstance to allocate available resources in the face of an unanticipated shortfall. These decisions must be accountable to general principles of ethical resource allocation, but do not follow a structured, systematic process (University of Toronto, 2005; AMA Council on Ethical and Judicial Affairs, 1995; Powell et al., 2008). Situational awareness is not available (i.e., the clinician making the decision is not in a position to manage resources or understand the magnitude of the event). Examples would include triage of multiple victims of an explosion to limited operating rooms immediately following the detonation. The goal is to minimize reactive triage decisions and assure those that are made are based on expert clinical judgment and ethical criteria.

TABLE 7

Characteristics of Reactive and Proactive Triage.

Proactive triage involves systematic decisions made by clinical or administrative personnel to a situation requiring resource triage where situational awareness is available and the decision-making is accountable to the incident management process. Examples would include prioritizing patients for evacuation from a facility and allocation of limited ventilators in a pandemic. Guidelines are available for some of these situations. For situations that lack specific guidance, appropriate subject-matter experts should weigh available information and make decisions consistent with principles of ethical resource allocation.

Prerequisite Command, Control, and Coordination Elements

The implementation of crisis standards of care and fair and equitable resource allocation requires attention to the core elements of incident management, including situational awareness, incident command, and adequate communication and coordination infrastructure and policies. Without this foundation, medical care will be inconsistent, and resources will not be optimally used (Hick et al., 2009).

Incident Management: Consistency, Coordination, and Communications

Incident management systems in the United States are based on a common framework called the National Incident Management System. The widely used Hospital Incident Response System is a NIMS-compliant incident management system modified for hospital applications (FEMA, 2009a; California Emergency Medical Services Authority, 2007). All healthcare facilities and entities must have a well-practiced incident management system and understand their plans for notification, activation, mobilization of resources, and continuity of operations.

Health and medical response is managed in the National Response Framework as outlined in Emergency Support Function (ESF) #8—Public Health and Medical Services (FEMA, 2009b; Courtney et al., 2009). At this time, ESF-8 does not have specific provisions for crisis standards of care. However, federal response partners should ensure the integration of relevant provisions. A system of a tiered response, ranging from healthcare management asset through federal responses, has been described by HHS and should be used by all hospitals and regional systems and are a core part of catastrophic response planning (Devereaux et al., 2008a; Phillips and Knebel, 2007; Courtney et al., 2009).

All healthcare systems must also understand how their incident management system interacts with that of jurisdictional emergency management and any coalition hospital response partners, including the process for obtaining assistance during an emergency (Figure 2).

FIGURE 2

Overview of relationships among agencies, committees, and groups NOTE: Depending on the organization of the state, the functional layout, details, and relationships among the units might vary.

Local/Regional Healthcare Coalitions

In many areas, regional healthcare coalitions exist that provide a common coordination point for hospital planning and response (Courtney et al., 2009; Phillips and Knebel, 2007; Hodge et al., 2009a). In certain environments, this coordination may be supplied by the state. Often, the coalition designates a Regional Medical Coordination Center (RMCC) function that coordinates hospital information and coordinates resource management during a major disaster (Burkle et al., 2007; Courtney et al., 2009; Phillips and Knebel, 2007). These coalitions may be within a jurisdiction, represent an entire jurisdiction, or overlap several jurisdictions or even states. Coalitions are generally organized around functional medical referral areas, however, as noted by Courtney et al. (2009):

The geographic boundaries of healthcare coalitions are highly variable, and the definition of community must remain flexible to incorporate local needs and realities. The essential feature is that every hospital in the chosen geographic area is included. In some places, the coalition may be composed of all hospitals and other members within a county or a city, while in others members may be from an entire state. In some small or low population density states, a single coalition may represent all hospitals and relevant partners in the entire state. In some large cities, the jurisdiction may be divided into [smaller] more manageable sub-municipal regions, so that a single city might have multiple coalitions. In many locations, coalitions cross jurisdictional borders and are not aligned with the normal geographic boundaries of all individual coalition members.

Healthcare coalitions should be designed to provide added administrative and logistical support to the many components of the health system that need to share limited resources or to transfer patients due to disaster situations. Notably, during a catastrophic disaster, reliance on the state or adjacent regions may become greater. Similar to traffic management or information technology networks, when one part of the system is overloaded, other parts of the system can help accommodate the load and maintain function. During a pandemic, limited or no “buffer” is available due to the pervasive nature of the epidemic, and the coalition function becomes coordination of consistent care across its members, rather than diffusion of demand across the system. A system of tiered response, from individual healthcare institutions as part of healthcare coalitions, coalitions as part of a jurisdiction(s) response, jurisdictional interface with the state, and the state’s interaction with federal response, have been described and should be used by all hospitals and regional systems as a core part of catastrophic response planning (Barbera and McIntyre, August 2004; Phillips and Knebel, 2007; Rubinson et al., 2008).

Coalitions streamline and facilitate resource allocation and policy coordination in disasters. The coalition must be integrated with key stakeholder agencies within ESF-8, including the broader “health system” (which may include clinics, long-term care facilities, behavioral health, and specialty resources, e.g., dialysis) as well as local and regional public health entities, emergency management entities, and emergency medical services. Often, these entities cross jurisdictions and are best coordinated using a Multi-Agency Coordination (MAC) (National Wildfire Coordinating Group, 1994). If no coalition is present in an area, hospitals must still integrate with the emergency management response (Figure 3).

FIGURE 3

HHS Medical Surge Capacity and Capability (MSCC) framework. NOTE: Emergency management program (EMP); emergency operations plan (EOP); public health (PH); emergency management (EM); healthcare organization (HCO); incident command system (ICS).

The MAC is the basis for establishing situational awareness and policy coordination across a given region, and incorporates data from the key participants, informing the decision-making process with regard to the transitions among conventional, contingency, and crisis care. It may also be delegated authority from participating agencies to manage scarce resources. The MAC and jurisdictional Emergency Operations Centers (EOCs) coordinate with the state EOC, though the degree of engagement and ability to make resource requests varies by state; emergency managers’ assistance must be engaged to assure the state’s system requirements are met by the MAC concept of operations.

Some areas of the United States have very robust and strong regional healthcare and emergency response coordination mechanisms that may be the decision point for crisis standards of care policy and resource allocation, while others will rely on the state for these functions.

Such coalition-building efforts can incorporate the presence of DoD military treatment facilities, of which there are more than 200 distributed on military bases across the United States. These facilities have assigned staff to attend to the emergency management requirements of their healthcare facilities, and most recently have designated a public health emergency officer on each of its bases to assist in the coordination of planning for a major public health emergency (Hachey, 2009). In addition, the VHA, with more than 150 medical centers across the country (some of which serve in the role as federal coordinating centers for the National Disaster Medical System), has championed the importance of its emergency management efforts, including plans to manage critical resource shortages (HHS, 2008; Franco et al., 2007; Bierenbaum et al., 2009; Department of Veterans Affairs, 2009; Sharpe, 2009). Commanders and directors of these facilities have authority to provide humanitar ian care in an emergency, and are often involved in community-based planning efforts.

Crisis Standards of Care Operations

When crisis care becomes necessary, a threshold has been crossed requiring that the affected institution(s) either quickly address the situation internally, or, more likely, appeal to partner facilities and agencies for assistance in either transferring patients to facilities with resources or bringing needed resources to the facility. If these strategies cannot be carried out, or if partner facilities are in the same situation (e.g., a pandemic influenza scenario), then systematic implementation of crisis standards of care at the state level may become necessary in order to codify and provide guidance for triage of life-sustaining interventions as well as to authorize care provided in non-traditional locations (alternate care facilities).

Because disaster incidents may have a wide-ranging impact on service delivery, a number of processes must occur, as described below.

Healthcare Facility Responsibilities

Though this section will emphasize emergency and hospital-based care, all healthcare facilities should have plans to preserve the acute care and other critical elements of their disaster services through elimination of certain usual services and curtailment of others. Taking an approach that incorporates “engineered failure” will ensure that those services that are absolutely essential will be maintained, at the expense of less pressing needs (Hick et al., 2007; ICDRM, 2009). For example, the delivery of dialysis care to patients with end-stage renal disease may be prioritized over out-patient elective surgery. A sample institutional process is outlined in Box 5 below.

BOX 5

Sample Institutional Process. Incident commander recognizes that systematic changes are or will be required to allocate scarce facility resources and that no regional resources are available to offset demand. Incident commander activates clinical care (more...)

Decision Tools and Resource Use Guidance

Decision tools are used by the triage team as a basis for, or to at least inform, triage decisions. Triage decision tools must be regionally consistent in a disaster event, highlighting the importance of the state as a source of guidance when possible. The healthcare coalition RMCC (or RDMAC, if established) can serve as the coordinator of policy, information, and process improvement. Intrastate consistency should be monitored by the SDMAC. The state department of health or governor should assure that the guidance they approve is consistent across state borders by consultation with adjacent state health departments (and EOCs during an event).

State guidance can also offer additional information about maximizing availability of the scarce resource to minimize impact on patients that may be specific to a resource or broader (Minnesota Department of Health, 2008). Decision tools and guidance should not be construed as to prevent reasonable consideration of other clinical factors that may weight a decision to provide or reallocate a scarce resource, but are issued to provide consistency and as much weight of evidence as possible to the decision-making process. This discussion provides a cross-section of available information that was the best available to the committee at the time of writing.

Although the most examined decision tools revolve around mechanical ventilation, guidance is also available for other core medical care components (medications, oxygen, etc.) and limited guidance is available for specific other resources (see Box 6) (Minnesota Department of Health, August 2008). Little guidance is available for the dispatch, EMS, home care, long-term care, and ambulatory care environments as part of the overall health system within a community. Though much of the core component guidance does apply, agencies and entities should examine potential scarce resources and outline coping strategies using base principles similar to those for hospital environments (Rubinson et al., 2008; ANA, 2008). None of the current systems or guidance was designed for pediatrics or other medical special needs patients, and this gap should be addressed by appropriate specialty expert groups. Finally, the needs of other vulnerable populations should also be kept in mind to ensure fairness in the system that is developed.

BOX 6

Select Specific Resource Issues. Note: synopsis and examples are not comprehensive, but suggest areas for state guidance and expert working group efforts. Blood products—The American Association of Blood Banks can facilitate blood delivery rapidly (more...)

Some literature is available to predict in-hospital requirements for critical care and to make general mortality predictions, and these may be useful when determining whether to hospitalize patients, send them home, or transfer them to an alternate care setting. However, these scores are not as useful in comparative prediction of mortality and are not precise, thus, the committee cannot recommend specific prognostic tools based on clinical assessment at this time (Talmor et al., 2007; Challen et al., 2007).

A concept originating in military triage which, though not a tool per se, may be used to weigh resource commitment is “minimum qualifications for survival” (MQS), which is the idea that one critically ill patient may consume the resources that could save several other patients, and may have their resource allocation reduced or withdrawn by establishing a ceiling on resources expended on a single patient (Christian et al., 2006). As an example, in military mass casualty experience described by Propper, 8 percent of patients consumed 43 percent of blood products used. In situations of resource shortages where the resource is titrated or dosed (medications, IV fluids, blood products, but not ventilators), the clinical care committee may wish to establish a ceiling on the amount of resources required in addition to changes to indications for treatment (Propper et al., 2009; Beekley et al., 2007; Eastridge et al., 2006; AMA, 2007).

Triage of limited mechanical ventilators may have to occur in pervasive events when no alternatives are available and temporizing therapies (e.g., bag-valve ventilation) cannot be implemented. Using the CDC’s Flu Surge 2.0 models for a severe pandemic (and assuming an 8-week pandemic wave, which is likely more compressed than what will be observed) suggests that at a busy, urban Level 1 trauma center, approximately 0.62 patients per hour may present during the peak weeks of the first wave with respiratory failure, necessitating ongoing monitoring and triage of resources to those with the best possible chance of survival (CDC, 2006). Those triaged to receive mechanical ventilation thus receive a therapeutic trial of ventilation. Predictions are applied to all ICU patients, not just incident-related patients. If the patient does not respond to an adequate trial, worsens, or another patient with a significantly better chance of benefit presents, the trial may be ended and the resource reallocated (Devereaux et al., 2008b; Hick and O’Laughlin, 2006; Christian et al., 2006; Hick et al., 2007; The Pandemic Influenza Ethics Initiative, 2008). Notably, the “therapeutic trial” may require days, as young, healthy individuals with severe pneumonia or respiratory distress syn drome may take many days to respond to treatment. All patients should be reassessed at least every 24 hours, however. As triage continues over days to weeks, the trend toward healthier patients on the available ventilators will likely reduce the degree of ventilator turnover compared to early in the triage process.

The impact of such decisions on providers and family, not to mention patients, cannot be understated and requires careful management of expectations on hospital admission as well as support and thoughtful transition plans as care is withdrawn to assure patient comfort and continued supportive care to the extent possible.

Due to the unique characteristics of ventilators (limited, expensive, technically complex resources that provides life-saving intervention and cannot be shared or titrated), much of the current decision tool efforts have centered around ventilator triage and critical care triage (Devereaux et al., 2008b; Christian et al., 2006; Hick et al., 2007; Hick and O’Laughlin, 2006). The decision tools generally are based on prognosis of the acute illness and any severe, underlying diseases that drastically limit life expectancy.

Guidelines for ventilator triage have already been adopted by several states and are in consideration by many others (Minnesota Department of Health, August 2008; The Utah Hospitals and Health Systems Association, January 2009; Colorado Department of Public Health and Environment, July 2009). These guidelines are generally based on several articles published in the past few years. Thus far, the Sequential Organ Failure Assessment (SOFA) score is used by all proposed systems as a core component (Vincent et al., 1996; Moreno et al., 1999; Vincent et al., 1998; Peres Bota et al., 2002; Pettila et al., 2002). SOFA uses clinical and some simple laboratory variables (PaO₂, bilirubin, creatinine) to predict outcome by assessing degree of organ system dysfunction and is one of the least complex and most predictive available metrics for prognosis prediction in critical care.

Some systems consider other factors such as expected duration of ventilation, underlying diseases, or duration of benefit (Minnesota Department of Health, August 2008; Devereaux et al., 2008b). Others incorporate exclusion criteria to varying degrees (The Pandemic Influenza Ethics Initiative, 2008, 2009; The Utah Hospitals and Health Systems Association, January 2009; Devereaux et al., 2008b; White et al., 2009; Christian et al., 2006; Hick and O’Laughlin, 2006). Incorporation of age as a specific variable has been proposed by one author (White et al., 2009).

Though the “fair innings” argument to allow ventilator allocation to younger patients is attractive at face value, age is not a medically useful predictor of outcome; use of age as a criterion in and of itself also raises ethical and legal concerns. Until society determines through public engagement that age-based triage (or other non-medical criteria such as functional capacity) is appropriate and defines an appropriate range, the committee recommends avoiding age-based criteria. Furthermore, the committee cautions against the prima facie use of DNR status as a decision tool, as underlying, life-limiting medical conditions should primarily be used as triage criteria rather than the fact that the patient has provided an advance directive.

The committee also notes that, although SOFA is useful to assign retrospective survival prediction, it was not designed as a prospective predictor of survival, and thus, differences in a single point on the SOFA scale are of unknown clinical significance for prediction of outcome. This should be considered, particularly when attempting any modification or extension of the SOFA scale beyond its initial construct that may further compromise its predictive value and when using systems that would assign or discontinue a resource based on a single-point change in the SOFA score.

SOFA has not been validated on a pediatric population. Although the principles of increasing mortality with increasing multi-organ dysfunction do apply, caution must be exercised when using SOFA to make anything but broad comparisons. Currently, predictive scoring systems for pediatrics (e.g. PRISM, P-MODS) are being considered for use in performing pediatric triage for ventilator allocation (Pollack et al., 1988; Graciano et al., 2005). However, at least one of these tools, PRISM, involves the evaluation of additional laboratory variables than those required for SOFA, and therefore might be more difficult to apply under conditions of crisis care. The other tool, P-MODS, evaluates parameters different than those used in SOFA scoring. The committee concludes that urgent recommendations from pediatric disaster groups and research are needed to address this gap. Adopters of decision tools should understand their limitations and scope and communicate issues of uncertainty to the triage team members.

The only process and triage system that is the output of an expert, specialty society working group with broad stakeholder input at this time is that of the American College of Chest Physicians (ACCP) (Devereaux et al., 2008b). The advantage of the ACCP process, though less specific than some systems, is that it considers duration of need and underlying disease in addition to the SOFA score acuity assessment. The basic triage process is outlined in Figure 4 and the exclusion criteria are described in Box 7, with additional supportive materials available in the original article. This process has informed most state guidance and other system guidance, including the VHA and other guidelines (Minnesota Department of Health, August 2008; The Pandemic Influenza Ethics Initiative, 2008, 2009; The Utah Hospitals and Health Systems Association, January 2009; Colorado Department of Public Health and Environment, July 2009).

FIGURE 4

Triage algorithm process. ^aExample exclusion criteria include severe, irreversible organ failure (CHF,liver, etc), severe neurologic compromise, extremely high or not improving SOFA scores, etc.

BOX 7

Exclusion Criteria Prompting Possible Reallocation of Life Saving Interventions. SOFA > 15 SOFA > 5 for >5 d, and with flat or rising trend

Critical care and ventilator allocation decision tools should be consistent with currently available evidence-based expert panel and national critical care guidelines, although modifications may be made to meet the specific needs of the state.

Of note, ventilators may not be the only relevant limitation to mechanical ventilation, as available staff, oxygen, and medication supply may not be able to support significantly more ventilators than the hospital normally uses due to design and supply limitations, thus, wholesale purchase of ventilators may not obviate the issue. Finally, decision tools may be supplemented by event-specific information (e.g., mortality data during a pandemic for particular underlying disease states or age ranges) or by supplemental prognostic information (e.g., as discussed in palliative care section). During an event such as a pandemic, federal guidance may be issued or epidemiologic information may be available that may affect state guidelines.

As evidence improves in triage science, modifications to these recommendations are likely. The state department of health or other appropriate office must maintain an advisory panel that can consider and incorporate necessary updates to this information prior to and during events and provide feedback on or assist with crisis clinical guidance development to ensure that the best available evidence is used should this type of triage be required. These state entities are encouraged to work with localities to ensure that local/regional coordination is occurring in real-time.

Recommendation 6: Ensure Consistency in Crisis Standards of Care Implementation

State departments of health, and other relevant state agencies, in partnership with localities should ensure consistent implementation of crisis standards of care in response to a disaster event. These efforts should include:

• Using “clinical care committees,” “triage teams,” and a state-level “disaster medical advisory committee(s)” that will evaluate evidence-based, peer-reviewed critical care and other decision tools and recommend and implement decision- making algorithms to be used when specific life-sustaining resources become scarce.
• Providing palliative care services for all patients, including provision of comfort, compassion, and maintenance of dignity.
• Mobilizing mental health resources to help communities—and providers themselves—to manage the effects of crisis standards of care by following a concept of operations developed for disasters;
• Developing specific response measures for vulnerable populations and those with special medical needs, including pediatrics, geriatrics, and persons with disabilities.
• Implementing robust situational awareness capabilities to allow for real-time information sharing across affected communities and with the “disaster medical advisory committee.”