Introduction
Modeling is a tool commonly used by preparedness agencies, including the Administration for Strategic Preparedness and Response (ASPR), to provide insights and data that improve understanding of high-consequence mass-casualty incidents (MCIs) such as nuclear detonation and support the development of response strategies, coordination of staff and resources, investment, development, and acquisition of medical supplies and medical countermeasures (MCMs).
Modeling indicates that a nuclear detonation over a large, densely populated U.S. city would have catastrophic consequences, potentially injuring hundreds of thousands of individuals, depending on the scenario.Reference Knebel, Coleman and Cliffer 1 These estimates include all people injured except immediate fatalities. It is important to realize that after a nuclear detonation, the need for resources will exceed supplies and therefore resources will become scarce and crisis standards of care will have to be implemented.Reference Cliffer, Wright and Yeddanapudi 2
Modeling results should not be viewed as a prediction of what would happen (i.e., reality) but rather viewed as rough estimates of what could happen under the assumptions and algorithms of the model. Modeling to support MCM investments should reflect the number of survivors who would likely receive medical care since funding and resources are limited. This paper demonstrates the use of a novel damage zone-based (DZ-based) sorting method to adjust rules used to determine which of the most seriously injured would be expected to die prior to rescue based on damage zone. This analysis is being used to inform the development of Biomedical Advanced Research and Development Authority’s (BARDA) trauma MCM portfolio.
HHS/ASPR’s Nuclear Detonation Model
A medical consequence model was developed for HHS/ASPR to provide estimates of the number of people injured by radiation, thermal radiation, and blast insults from a nuclear detonation.Reference Cliffer, Wright and Yeddanapudi 2 Results of this ASPR Model also include the quantity of various MCMs required to treat the surviving injured. Multiple scenarios have been modeled using different U.S. cities, nuclear yields, heights of detonation, and seasonal wind patterns. Within ASPR, BARDA has been using the ASPR Model discussed here over the last 15 years to support its MCM preparedness program.
The ASPR Model bases its estimates of injured people on population and nuclear weapon effects (i.e., airblast overpressure, ionizing and thermal radiation) provided at gridded cells centered at the point of detonation. Weapon effect levels decrease with distance, so the heaviest structural damage and highest immediate fatality counts will occur closest to ground zero. The location of people relative to buildings (e.g., inside a building near windows, at the building core, outside in the open, or outside near a building) affects injury mechanisms and severity since people at the core of a building would be protected from glass injury and thermal radiation while people outside would not receive shielding from radiation exposure. The nuclear weapon effect levels at each grid cell are compared with published injury thresholds to determine if an injury occurs. Thermal burn injuries are categorized by degree of burn (using assumed thermal burn thresholds) 3 and extent, which is calculated using a simple cylindrical model with a protection factor applied to clothed areas of the body.Reference Levin 4 Radiation injuries are characterized by whole-body radiation doses considering prompt and residual radiation (fallout) and shielding from various building types.
The ASPR Model includes three mechanisms of blast injury: flying glass shards due to broken windows (based on shock tube experiments on animals),Reference Fletcher, Richmond and Yelverton 5 building collapse (with injury severity distributions based on earthquake data),Reference Rich, Callahan and Hassol 6 and high-velocity physical body displacement for people outside at the time of detonation (based on tumbling and impact with a rigid object). 3 Trauma injuries are categorized as mild (injury severity score [ISS] range 1-8), moderate (ISS range 9-15), severe (ISS range 16-24), or very severe (ISS > 24). 7
In addition to trauma injuries, blast forces also cause structural damage, resulting in collapsed buildings and rubble in the streets. Along with lingering radiation, this structural damage will result in delays in rescue operations, delivery of medical supplies, and transportation of patients from the point of injury to operational medical treatment facilities.
As noted by Knebel and colleagues,Reference Knebel, Coleman and Cliffer 1 previous modeling results estimate that the number of injured survivors could be in the thousands to hundreds of thousands depending on the scenario (yield, height of detonation, city, winds). These estimates can be improved to better represent who may receive interventions, whether treatment or palliative, by considering DZs, time to rescue, and scarce resource decision-making.
Nuclear Damage Zones
After a nuclear detonation, disaster response operations, including search and rescue, would be organized according to three physical DZs centered at ground zero, as illustrated in Figure 1. Starting from ground zero and working outward, 3 concentric circular DZs are defined based on peak overpressure: severe DZ (red), moderate DZ (orange), and light DZ (yellow).
Visualization of nuclear blast damage zones.
The extent of the DZs depends on the weapon yield and detonation height. Damage zones are used to categorize the extent of physical building and infrastructure damage as indicated in Figure 2. 8
Blast damage zones including observable features.
Figure 2. Long description
The top half of the diagram shows a city map with an X at the bottom-left corner marking a Notional 10 k T Detonation. A color-coded gradient radiates from this point across a grid measured in miles on both the x and y axes.
From the detonation point moving outward:
1. Severe Damage Zone S D Z. Colored dark red, extending to approximately 0.5 miles. Features include nearly all buildings destroyed, few survivors, and hazardous outdoor radiation levels.
2. Moderate Damage Zone M D Z. Colored orange, extending from 0.5 to approximately 1.5 miles. Features include significant building damage and rubble, light buildings destroyed, blown out interiors of larger buildings, utility poles downed, fires, and a significant number of major injuries where early medical assistance can improve survival.
3. Light Damage Zone L D Z. Colored yellow, extending from 1.5 to approximately 3.0 miles. Features include nearly all windows shattered, damage to building facades, and mostly minor injuries from flying glass and debris that are highly survivable without immediate medical care.
4. Beyond L D Z. Colored white, extending beyond 3.0 miles. Features include windows broken and doors blown in primarily on the side facing the blast, and few, generally minor, injuries.
A legend at the bottom defines Structural Damage symbols used on the map:
- Red diamond: Collapsed Building
- Purple diamond: Severely Damaged Building
- Blue square: Moderately Damaged Building
- Tan square: Shattered Glass Injuries
- Black square: Broken Windows
Damage zones can also be used to estimate when first responders can begin rescue operations. 8 , 9 Within the severe DZ, collapsed buildings and rubble will make streets impassable and hinder prompt rescue, with very high radiation levels also delaying rescue efforts. Rescue efforts in the severe DZ are not a priority in the first several days after detonation until the fallout decays and significant progress has been made in the moderate DZ. Within the moderate DZ, there will be rubble in the streets, although less than in the severe DZ, allowing for better access by search and rescue teams. Except in the dangerous radiation zone, radiation levels will quickly subside to levels acceptable for first responders, enabling rescue efforts to begin. The initial focus of rescue operations in the light DZ will be helping severely injured people.
Triage and Expectant Categorization Rules
Detonation of a nuclear device in a populated area would cause hospitals to become overwhelmed and resources to become scarce.Reference Knebel, Coleman and Cliffer 1 , Reference Casagrande, Wills and Kramer 10 TriageFootnote a becomes an important way to prioritize patient care so that available resources can provide the “greatest good for the greatest number.”Reference Bazyar, Farrokhi and Salari 11 , Reference Coleman, Weinstock and Casagrande 12 Triage involves categorizing individual patients into different groups (immediate, delayed, minor, and deceased or expectant) based on the type and severity of their injuries and the availability of resources.Reference Iserson and Moskop 13 Expectants are defined as people “with serious or multiple injuries requiring intensive treatment, or with a poor chance of survival. These patients receive appropriate supportive treatment compatible with resources, which will include large doses of analgesics as applicable.” 14
Clinical triage protocols for MCIs typically focus on parameters such as respiratory rate, vital signs, and motor responses of the individual. These parameters are not amenable to disaster preparedness models that track groups of people and not individuals.Reference Hick, Hanfling and Cantrill 15 , Reference Sullivan, Prasanna and Grace 16 Other triage methods, such as Coleman-Weinstock and DOW/NATO triage, are compatible with modeling because they consider severity levels of radiation exposure, thermal burns, and trauma injuries, rather than an individual’s clinical parameters. The Coleman-Weinstock triage method provides recommendations for treatment priority during the first 4 days after a nuclear detonation when crisis standards of care would apply.Reference Coleman, Weinstock and Casagrande 12 It is limited in operational usage by its assumption of immediate availability of biodosimetry for estimating radiation dose to categorize patients deemed expectant.Reference Sullivan, Prasanna and Grace 16 The DOW/NATO triage method acknowledges this limitation and evaluates trauma and burn injuries independent of radiation exposure. It does not rely on radiation dose estimates to categorize expectants except at very high doses that induce readily observable symptoms such as extreme nervousness and confusion, severe nausea, early and/or severe vomiting, watery diarrhea, and loss of consciousness. 14 , 17 The choice of DOW/NATO triage reflects a more realistic ability to assess radiation exposure. Use of Coleman-Weinstock triage would result in many more people being categorized as expectant from radiation.
Methods
We conducted a study investigating which specific trauma injuries might be expected to occur as a result of a nuclear detonation. We expanded the ASPR model beyond the existing injury severity levels (minor, moderate, severe, and very severe) to include estimates of specific types of trauma injuries that would occur from a nuclear detonation and need targeted medical interventions. As part of this modeling, we developed a new population sorting method called damage zone-based (DZ-based) sorting, which introduces the concept of unrescuable. Unrescuable individuals are defined as people who won’t survive long enough to be rescued. We then examined the impact of the DZ-based sorting method on the distribution of treated injury severities, estimates of trauma-injured people to be treated, and the estimates of trauma injuries to be treated.
Population Sorting Methods
The ASPR Model estimates the number of people to be treated for their injuries and has historically relied on triage protocols to estimate expectant individuals. Recognizing that many would die prior to being rescued and thus would not require palliative care, we defined a new population group called unrescuable individuals, to be used in addition to the 2 population groups we were already modelling (treated individuals and expectant individuals). We defined 3 population sorting methods: (1) no sorting, (2) sorting based on DOW/NATO expectant rules, and (3) DZ-based sorting.
The simplest population sorting method is to categorize everyone as treatable. We call this “no sorting.” This overestimates the number of people to be treated in a resource-constrained setting since many people would have extensive injuries requiring more resources than are available, and it, by definition, does not break out expectant or unrescuable individuals. It is a fictional construct for the purposes of modeling, assuming that all injured people are treated, ignoring the limitations on resources. These results can be used for heuristic purposes to set a boundary on requirements, but they are not realistic.
Sorting solely based on DOW/NATO expectant rules refines the estimates of the number of people to be treated and also estimates the number of people that will need palliative care rather than injury treatment (i.e., expectant individuals). This improves the accuracy of our estimates of MCM quantities required to treat the injured and to provide palliative care to expectant individuals. Sorting based on expectant rules still overestimates the number of survivors, especially in the severe damage zone. It does not consider the time that the injured must wait before they can be rescued and receive medical attention. During this time, injuries would worsen, and many expectants and some potentially savable people will die.
Recognizing that many people in areas with the heaviest damage would die prior to being rescued, we further refined our estimates using a DZ-specific application of expectant rules, which we called “DZ-based population sorting.” It categorizes the trauma-injured population as treat, expectant, or unrescuable. Unrescuable individuals wouldn’t survive long enough to be rescued, because it would take many days before responders reach them. These individuals would not be able to receive treatment MCMs or palliative care. The logistics of transportation of survivors from point of injury to a medical care facility and the resultant impact of associated delays on survivability were outside the scope of this analysis, although palliative care applied in the field would alleviate some transportation need.
The population sorting rules for all 3 sorting methods are shown in Table 1. The no-sorting method categorizes everyone as treatable (Table 1A). Sorting based on DOW/NATO expectant rules (Table 1B) categorizes individuals as treatable or expectant (purple) depending on their injury severity. DZ-based population sorting accounts for damage zones in addition to injury severity (Table 1C), and categorizes individuals as treatable, expectant (purple), or unrescuable (black).
Population sorting rules
Table 1. Long description
The table is divided into three sections: A, B, and C. Each section uses Damage Zone (D Z) as rows (Light D Z, Moderate D Z, Severe D Z) and Trauma injury severity as columns (Very Severe, Severe, Moderate, Minor).
Section A: No Sorting. All combinations of Damage Zone and Trauma injury severity result in the action Treat.
Section B: Sorting based on D O W or NATO Expectant Rules.
- For all Damage Zones (Light, Moderate, Severe), Very Severe and Severe trauma are labeled Expectant.
- Moderate trauma is labeled Treat unless Expectant (asterisk).
- Minor trauma is labeled Treat unless Expectant (double asterisk).
Section C: D Z-Based Sorting.
- Light D Z: All trauma levels are labeled Treat.
- Moderate D Z: Very Severe and Severe trauma are labeled Expectant. Moderate trauma is Treat unless Expectant (asterisk). Minor trauma is Treat unless Expectant (double asterisk).
- Severe D Z: Very Severe and Severe trauma are labeled Unrescuable. Moderate trauma is 57 percent Unrescuable and 43 percent Treat unless Expectant (asterisk). Minor trauma is Treat unless Expectant (double asterisk).
Footnotes:
- Asterisk: Moderate trauma-injured people are expectant if radiation exposure is at least 20 G y or second-degree burns cover at least 30 percent of the body.
- Double asterisk: Minor trauma-injured people are expectant if radiation exposure is at least 20 G y or second-degree burns cover at least 40 percent of the body.
* Moderate trauma-injured people are considered expectant if their radiation exposure is at least 20 Gy or they have second-degree burns covering at least 30% of their body.
** Minor trauma-injured people are considered expectant if their radiation exposure is at least 20 Gy or they have second-degree burns covering at least 40% of their body.
There are several reasons why DZ-based population sorting categorizes everyone in the light DZ, including all severe and very severe trauma-injured people, as treatable. The initial focus of rescue operations in the light DZ would be helping severely injured individuals, 9 and transportation in the light DZ would be less hampered by rubble and collapsed buildings. Individuals injured in the light DZ would likely be rescued promptly, allowing them to receive medical treatment in time to be effective. Discussions with in-house trauma care subject matter experts indicated that hospital staff would medically triage patients, prioritizing the most seriously injured until resources become scarce and crisis standards of care are implemented, which would not occur immediately.
In the moderate DZ, DZ-based population sorting follows DOW/NATO expectant rules. Response guidance suggests rescue operations could begin in the moderate DZ as soon as day 2, by which time hospitals could be assumed to have implemented their crisis standards of care. Populations with severe and very severe trauma injuries are sorted as expectants, while populations with minor and moderate trauma are sorted as treatable unless they suffer from expectant-level burns or radiation exposures.
In the severe DZ, DZ-based population sorting categorizes individuals with severe and very severe trauma as unrescuable individuals since rescue operations won’t begin in the severe DZ until day 3 or later, and those individuals would not survive without care long enough to be rescued. Individuals with minor trauma are categorized as treatable unless they suffer from expectant-level burns or radiation exposures, assuming they are not inextricably trapped in the rubble. Individuals with moderate trauma are categorized as treatable, expectant, or unrescuable, assuming they are not inextricably trapped in the rubble. The fraction categorized as unrescuable (57%) is based on time-to-death estimatesReference Casagrande, Wills and Kramer 10 and in-house trauma care subject matter expertise using response operations guidance for first responders entering the severe DZ. 8 , 9
Modeling of Specific Blast Trauma Injury Types
We also expanded the ASPR Model to estimate numbers of specific types of blast trauma injuries beyond the existing injury severity levels (minor, moderate, severe, and very severe).Reference Baker, O’Neill and Haddon 18 This blast trauma injury modeling effort involved a literature review in conjunction with guidance from subject matter experts to formulate a portfolio-relevant list of specific trauma injuries including but not limited to lacerations, fractures, crush injury, and hemorrhage. Modeling specific types of trauma injuries caused by a nuclear detonation is challenging due to the very limited availability of real-world data since the most seriously injured people in Hiroshima and Nagasaki perished before they could be hospitalized. Published analyses of hospitalized injuries in the aftermath of MCIs enabled us to develop an incidence of occurrence for each specific trauma injury type and severity level as shown in Table 2.
Sources used to define injury incidences
Table 2. Long description
The table is structured with injury types in the first column and severity categories across the top. Severity is divided into Minor, Penetrating (Moderate, Severe, Very Severe), and Blunt (Moderate, Severe).
* Soft tissue injuries, Fractures/dislocations, and Head injuries: All use Ref. 19 for Minor and Penetrating categories. For Blunt trauma, Soft tissue uses Ref. 21, while Fractures and Head injuries use Ref. 22 and Ref. 23 respectively.
* Eye injuries: Minor uses Ref. 20; Penetrating uses Ref. 19. Blunt trauma is noted as applicable only to penetrating trauma.
* Crush injuries: Minor and all Penetrating levels use Ref. 20. Blunt trauma uses Ref. 22.
* Severe lacerations: Not considered minor or moderate penetrating; Severe and Very Severe penetrating use Ref. 19. Blunt trauma uses Ref. 22.
* Compartment syndrome and Crush syndrome: Marked as applicable only to blunt trauma, using Ref. 22 and Ref. 24 respectively.
* Compressible hemorrhage to the extremities and Non-compressible torso hemorrhage: Not considered minor or moderate; Severe and Very Severe penetrating use Ref. 25 and Ref. 26 respectively. Blunt severe trauma uses the same references.
* Junctional hemorrhage: Applicable only to very severe penetrating injuries using Ref. 27.
A reference key at the bottom identifies the sources:
19: Shariat 1998 (Oklahoma City bombing)
20: M M W R 2002 (World Trade Center attack)
21: Prabowo 2018 (Earthquake)
22: MacKenzie 2017 (Earthquake)
23: DelPapa 2019 (Earthquake)
24: Naghii 2005 (Earthquake)
25: Heldenberg 2016 (Terror-related explosions)
26: Stannard 2013 (Battlefield casualties)
27: Kelly 2008 (Battlefield casualties)
Minor injuries include soft tissue injuries (lacerations, contusions, sprains, and strains), fractures and dislocations, head injuries, eye injuries, and crush injuries. Minor injuries do not require hospitalization. Injury incidences for minor injuries were estimated using hospital reports of people who were treated and released after the Oklahoma City bombingReference Shariat, Mallonee and Stidham 19 and the World Trade Center attack. 20
Moderate, severe, and very severe injuries are serious enough to require hospitalization; injury incidences were estimated using hospital reports of people who were hospitalized in the aftermath of MCIs such as earthquakes,Reference Prabowo, Kamal and Latief 21 -Reference Naghii 24 terrorist bombings,Reference Shariat, Mallonee and Stidham 19 , Reference Heldenberg, Givon and Simon 25 and the World Trade Center attack. 20 Battlefield casualty data were used when data were lacking for civilian incidents.Footnote b , Reference Heldenberg, Givon and Simon 25 , Reference Stannard, Morrison and Scott 26
Injury incidences were determined for 3 kinds of hemorrhage. Compressible hemorrhage is bleeding that can be stopped with direct pressure or tourniquets, and, in this study, specifically means bleeding from the extremities. Non-compressible torso hemorrhage, on the other hand, is bleeding that cannot be stopped with direct pressure, tourniquets, or junctional pressure devices. Junctional hemorrhage is a third kind of hemorrhage, a term most often encountered in military medical literature, that is “non-tourniquet-able but compressible”Reference Kelly, Ritenour and McLaughlin 27 or hemorrhage occurring at the axillary, neck, or groin areas.
The blast trauma injury incidences determined from these sources, together with the estimates of the number of people injured from the ASPR Model, were used to estimate the number of specific trauma injuries. The ASPR Model predicts that people will likely suffer more than one trauma injury, distinguishing counts of the injured from counts of trauma injuries.
Quantifying the Effect of Population Sorting
To assess the impact of population sorting on our modeling results, we ran the ASPR Model with all 3 population sorting methods for 3 different scenarios, including 1-kt and 10-kt ground burst nuclear detonations and a 10-kt nuclear airburst in a large metropolitan area. The airburst scenario allowed us to assess the changes in the number of people with burns, radiation exposure, and trauma injuries between the scenarios. All 3 scenarios used identical ground zero location, wind pattern, and timing of the incident. The no-sorting method was used as the standard against which the results of different sorting methods are demonstrated. Results do not include immediate fatalities or people injured by radiation or burn only. The effect of population sorting was examined using 3 metrics.
The effect of population sorting on the distribution of trauma-injured people by injury severity and population category is calculated as follows.
where x is minor, moderate, severe, very severe, treat, expectant, or unrescuable.
The effect of population sorting on the percentage of people to be treated for trauma injuries is calculated as follows.
The effect of population sorting on the percentage of trauma injuries to be treated is calculated as follows.
Results
Effect of Population Sorting on the Distribution of Trauma-Injured People
The distribution of trauma-injured people by injury severity and population category is presented in Table 3 for 3 scenarios. Under DZ-based sorting, the severely trauma-injured population categorized as treatable represents only a small fraction of the total (4% to 9% depending on scenario), with the very severe trauma-injured categorized as treatable representing an even smaller fraction (0.2% to 0.5%). Compare this with the 0% of severe and very severe trauma population treated using sorting based on DOW/NATO expectant rules. Overall, DZ-based sorting results in a higher proportion of trauma-injured population being treated than sorting based on DOW/NATO expectant rules for all 3 scenarios. Expectants under DZ-based sorting make up about half the fraction of trauma-injured compared with sorting based on DOW/NATO expectant rules for all scenarios (purple column).
Distribution of trauma-injured people by injury severity and population sorting category
Table 3. Long description
The table is divided into three main sections based on burst type.
1. 1 k t Ground Burst section
* No sorting: Minor 22%, Moderate 51%, Severe 18%, Very Severe 9%. Sorting: Treat 100%.
* D O W / NATO expectant rules: Minor 20%, Moderate 46%, Severe 0%, Very Severe 0%. Sorting: Treat 66%, Expectant 34%.
* D Z-based sorting: Minor 21%, Moderate 47%, Severe 9%, Very Severe 0.5%. Sorting: Treat 77%, Expectant 15%, Unrescuable 8%.
2. 10 k t Ground Burst section
* No sorting: Minor 20%, Moderate 31%, Severe 29%, Very Severe 19%. Sorting: Treat 100%.
* D O W / NATO expectant rules: Minor 18%, Moderate 26%, Severe 0%, Very Severe 0%. Sorting: Treat 44%, Expectant 56%.
* D Z-based sorting: Minor 18%, Moderate 26%, Severe 6%, Very Severe 0.3%. Sorting: Treat 55%, Expectant 34%, Unrescuable 16%.
3. 10 k t Airburst section
* No sorting: Minor 18%, Moderate 26%, Severe 28%, Very Severe 28%. Sorting: Treat 100%.
* D O W / NATO expectant rules: Minor 18%, Moderate 26%, Severe 0%, Very Severe 0%. Sorting: Treat 44%, Expectant 56%.
* D Z-based sorting: Minor 18%, Moderate 24%, Severe 4%, Very Severe 0.2%. Sorting: Treat 46%, Expectant 24%, Unrescuable 30%.
Figure 3 visually displays the injury severity of trauma-injured people for the 1-kt ground burst scenario using the data in Table 3. It illustrates the recategorization of people from no-sorting results (top) to sorting based on DOW/NATO expectant rules (middle), to DZ-based sorting (bottom). This format makes it easy to see the impact of population sorting.
Distributions of trauma-injured people by injury severity, 1 kt ground burst.
Figure 3. Long description
The table consists of four columns: Population Sorting Method, Distribution, Trauma Injury Severity, and percentage of Trauma Injured People.
Row 1: No Sorting.
* Pie chart shows four segments: yellow, green, orange, and red.
* Data table: Minor (green) 22 percent, Moderate (yellow) 51 percent, Severe (orange) 18 percent, Very Severe (red) 9 percent. Expectant and Unrescuable are blank.
Row 2: D O W forward slash NATO expectant rules.
* Pie chart shows three segments: yellow, green, and pink.
* Data table: Minor (green) 20 percent, Moderate (yellow) 46 percent, Severe 0 percent, Very Severe 0 percent, Expectant (pink) 34 percent. Unrescuable is blank.
Row 3: D Z-based sorting.
* Pie chart shows six segments: yellow, green, pink, black, orange, and a sliver of red.
* Data table: Minor (green) 21 percent, Moderate (yellow) 47 percent, Severe (orange) 9 percent, Very Severe (red) 0.5 percent, Expectant (pink) 15 percent, Unrescuable (black) 8 percent.
Effect of Population Sorting on the Percentage of Trauma-Injured People to Be Treated
The percentage of trauma-injured people to be treated for each trauma injury severity is provided for 3 nuclear scenarios in Table 4 under population sorting. Nearly the entire population with minor and moderate trauma injuries (92% to 100%) is categorized as treated under population sorting based on DOW/NATO expectant rules and DZ-based sorting. DZ-based sorting results in between 16% and 51% of the severe trauma and 1% and 6% of the very severe trauma population categorized as treated, depending on the scenario. Overall, DZ-based sorting categorizes a larger proportion of trauma-injured people as to be treated than sorting based on DOW/NATO expectant rules for all 3 scenarios.
Percent of trauma-injured people treated under population sorting
Table 4. Long description
The table is divided into three main sections based on detonation type. Each section compares D O W / NATO expectant rules against D Z-based sorting across five severity columns: Minor, Moderate, Severe, Very Severe, and Overall.
* 1 k t Ground Burst:
- D O W / NATO expectant rules: 99 percent Minor, 98 percent Moderate, 0 percent Severe, 0 percent Very Severe, 72 percent Overall.
- D Z-based sorting: 99 percent Minor, 98 percent Moderate, 51 percent Severe, 6 percent Very Severe, 81 percent Overall.
* 10 k t Ground Burst:
- D O W / NATO expectant rules: 97 percent Minor, 93 percent Moderate, 0 percent Severe, 0 percent Very Severe, 49 percent Overall.
- D Z-based sorting: 97 percent Minor, 93 percent Moderate, 21 percent Severe, 2 percent Very Severe, 55 percent Overall.
* 10 k t Airburst:
- D O W / NATO expectant rules: 100 percent Minor, 99 percent Moderate, 0 percent Severe, 0 percent Very Severe, 44 percent Overall.
- D Z-based sorting: 100 percent Minor, 92 percent Moderate, 16 percent Severe, 1 percent Very Severe, 47 percent Overall.
Effect of Population Sorting on the Percentage of Trauma Injuries to Be Treated
To further examine the impact of population sorting, we evaluated the percentage of treatable trauma injuries under sorting compared with under no sorting for each scenario (Table 5). DZ-based sorting categorizes a higher proportion of each specific trauma injury and of the overall trauma injuries as treatable than does sorting based on DOW/NATO expectant rules. The specific injury types which are considered more serious (i.e., severe lacerations and hemorrhage) have lower treated percentages than injury types which are considered less serious (i.e., soft tissue injuries, fractures, head injuries, eye injuries, and crush injuries) for all scenarios under sorting based on DOW/NATO expectant rules and DZ-based sorting. This follows directly from the population sorting rules where severe and very severely injured in the moderate and severe DZ are categorized as expectant or unrescuable. Unlike sorting based on DOW/NATO expectant rules, DZ-based sorting maintains the quantification of these more serious injury types by categorizing the severe and very severely injured from the light DZ as to be treated. Table 5 illustrates how the choice of population sorting method can make a large difference in requirements setting for more serious injury types.
Percent of specific trauma injuries treated under population sorting
Table 5. Long description
The table is organized with Trauma injury types in the first column and three main scenario headers: 1 k t ground burst, 10 k t ground burst, and 10 k t airburst. Each scenario is split into two columns: D O W forward slash NATO expectant rules and D Z-based sorting.
* Soft tissue injuries: 1 k t ground (72 percent vs 81 percent), 10 k t ground (48 percent vs 55 percent), 10 k t airburst (42 percent vs 46 percent).
* Fractures forward slash dislocations: 1 k t ground (65 percent vs 77 percent), 10 k t ground (38 percent vs 46 percent), 10 k t airburst (34 percent vs 37 percent).
* Head injuries: 1 k t ground (67 percent vs 79 percent), 10 k t ground (41 percent vs 49 percent), 10 k t airburst (34 percent vs 39 percent).
* Eye injuries: 1 k t ground (72 percent vs 82 percent), 10 k t ground (48 percent vs 55 percent), 10 k t airburst (41 percent vs 46 percent).
* Crush injuries: 1 k t ground (61 percent vs 70 percent), 10 k t ground (32 percent vs 39 percent), 10 k t airburst (42 percent vs 36 percent).
* Severe lacerations: 1 k t ground (0 percent vs 35 percent), 10 k t ground (0 percent vs 13 percent), 10 k t airburst (2 percent vs 9 percent).
* Non-compressible torso hemorrhage: 1 k t ground (0 percent vs 36 percent), 10 k t ground (0 percent vs 13 percent), 10 k t airburst (0 percent vs 8 percent).
* Compressible hemorrhage: 1 k t ground (0 percent vs 36 percent), 10 k t ground (0 percent vs 13 percent), 10 k t airburst (0 percent vs 8 percent).
* Junctional hemorrhage: 1 k t ground (0 percent vs 6 percent), 10 k t ground (0 percent vs 2 percent), 10 k t airburst (0 percent vs 1 percent).
* All injury types: 1 k t ground (64 percent vs 76 percent), 10 k t ground (40 percent vs 47 percent), 10 k t airburst (34 percent vs 38 percent).
Discussion
Models are designed to be useful for their purpose, and they must balance complexity with ease of use, making assumptions to deal with uncertainties and concepts that are outside their scope. Modelers make many assumptions such as injury thresholds, location of people, building types, injury severity distributions, treatment protocols, and triage expectant rules. This enables the estimation of the numbers of injured and expectant people, and the quantities of required MCMs. For this analysis, we made additional assumptions about time to rescue, survivability without treatment, the applicability of MCI injury data sets to a nuclear detonation scenario, and the incidence of specific types of trauma injuries. This helps estimate the number of specific trauma injuries that would need to be treated after a nuclear detonation.
Compared with no sorting and sorting based on DOW/NATO expectant rules, DZ-based sorting incorporates additional realism into our modeling results in 2 ways: (1) by introducing the concept of unrescuable individuals and (2) by acknowledging that early hospital arrivals with severe and very severe injuries would be treated until crisis standards of care are established. Recognizing that many individuals, particularly in the severe DZ, won’t survive long enough to be rescued drives down the estimates of expectants compared with sorting based on DOW/NATO expectant rules. Removing the unrescuable population from the expectant and treatable population is valuable for an analysis that informs investment in MCMs and palliative care.
The implementation of crisis standards of care will not happen immediately, which increases the number of severe and very severely injured people categorized as treatable compared with sorting based on DOW/NATO expectant rules. DZ-based sorting categorizes a larger proportion of trauma-injured people and a larger proportion of trauma injuries as to be treated than sorting based on DOW/NATO expectant rules for all 3 scenarios. If population sorting based on DOW/NATO expectant rules was used for medical planning, our model would estimate that no severe lacerations or hemorrhages would need to be treated in a nuclear scenario, and thus no investment would be made in the MCMs used to treat those injury types. DZ-based sorting gives these types of injury representation even if our modeling assumptions about treatment of the most seriously injured are overly optimistic. Future analysis could examine the practicability of treating severe and very severely injured in an MCI and the timing of the implementation of crisis standards of care to add further realism to these estimates.
Previous modeling incorporated expectant rules to reflect the scarce resources environment that a nuclear detonation would cause, and it is prudent to recognize that practical limitations would likely limit the possibility that everyone injured may receive medical intervention. In the process of expanding the ASPR Model to include trauma injury types, we realized we needed to apply a more nuanced approach with the expectant rules. To stay within the framework of our scope and keep the approach simple, expectant rules were modified for each damage zone. Our DZ-based sorting method was designed to serve the purpose of our analysis by incorporating a reduction in injured counts consistent with who may potentially receive medical intervention while maintaining the inclusion of some vital interventions that may be otherwise lost if injured counts are reduced too drastically.
Conclusion
The approach advocated by this paper is the application of a damage zone framework and expectant rules to provide more realistic estimates of the number of trauma-injured people to be treated compared with total injured counts such as cited by Knebel.Reference Knebel, Coleman and Cliffer 1 The choice of expectant rules and modifications within the damage zones should be consistent with the needs of the modeling and analysis. The DZ-based population sorting method not only allows for some consideration of on-the-ground realities but also incorporates the benefit of potential medical decision practices. As such, the DZ-based population sorting method allows for more realistic projection of medical needs for the provision of treatments and palliative care and could support strategies for the MCM development and procurement.
Author contribution
A.V. and N.I. conceived and designed the project. S.W., N.Y., and A.V. collected and reviewed literature. S.W. and N.Y. developed the statistical model and analyzed the data. S.W., N.Y., and A.V. wrote the first draft of the manuscript. All authors contributed to data interpretation, revised the manuscript critically, and approved the final submission.
Competing interests
None.
Disclaimer
The views, opinions, and/or findings expressed are those of the authors and should not be interpreted as representing the official views or policies of the U.S. Department of Health and Human Services or the U.S. government.
Open access
