Increased Incidence Of Inhalation Burn Injury During The Covid-19 Pandemic: A National Database Study

Abstract The COVID-19 pandemic has forced many Americans to adapt their daily routines. In 2020, there was a significant increase in house fires according to the National Fire Prevention Association (NFPA). The objective of this study was to characterize the changes in suspected smoke inhalations during the first year of the pandemic in the National Emergency Medical Services Information System (NEMSIS). The NEMSIS database was queried for all EMS transports captured between 2017-2020. Differences in the incidences of suspected smoke inhalations and fire dispatches in 2020 were estimated using Poisson regression models. There was a 13.4% increase in the incidence of fire dispatches and a 15% increase in suspected smoke inhalations transported in 2020 compared to the previous 3 years. The IRR of both fire dispatches (1.271; 95% CI: 1.254-1.288; p<0.001) and suspected smoke inhalation (1.152; 95% CI: 1.070-1.241; p<0.001) was significantly elevated in 2020. The increases in fire dispatches and suspected smoke inhalations observed in the NEMSIS database are in concordance with other literature indicating the increase in fire incidence and morbidity observed during the pandemic. These results should inform fire prevention outreach efforts and resource allocation in burn centers in the event of future pandemic.


INTRODUCTION
The COVID-19 pandemic has contributed to significant changes in the daily lives of Americans [1]. There have also been changes seen in patterns of traumatic injuries, delays in cancer diagnoses, and mental health [2][3][4][5]. Notably, the National Fire Prevention Association (NFPA) 2020 annual report noted the number of house fires increased significantly during the first year of the pandemic. The report, which describes a 5% increase in residential fires, suggests that changes in daily activities contributed to this increased fire incidence; many Americans are now spending more time at home, commuting less, and cooking more [6]. While the report also cites a decrease in per capita deaths due to fire, some studies in the literature have reported an increase in severity and morbidity during the pandemic. Yamamoto, et. al found an increase in in-hospital mortality, inhalation injuries, and total body surface area [7].
Inhalation injury (IH) is a major determinant of mortality from severe burn injuries, and represents increased morbidity and injury severity in victims of fires [8]. Management is challenging for healthcare providers and it creates a significant burden of disease to burn centers [8,9]. This injury pattern frequently presents in victims of house fires, as structural fires cause extreme air temperatures and generate smoke containing various chemicals, products of incomplete combustion, and aerosolized debris; fire victims that are confined to a closed space will have perpetuation of these injuries [10].
Currently, there are limited data examining the nationwide incidence of injuries from house fires during the pandemic period. We examined a national prehospital database in an effort to better characterize the changes in these injury patterns using suspected smoke inhalation (SSI) as a surrogate of morbidity from fires. We hypothesized that there was an increase in suspected smoke inhalation in 2020 and fire dispatches during the first year of the COVID-19 pandemic.
A c c e p t e d M a n u s c r i p t

Study design, data source and population
An institutional review board approval was obtained to perform a retrospective study using the National Emergency Medical Information System (NEMSIS). The NEMSIS is a convenience sample of emergency 911 activations nationwide, maintained by the National Highway Traffic Safety Administration (NHTSA) aimed at improving patient care through standardization, aggregation, and utilization of point of care EMS data at a local, state, and national level. It contains information on response characteristics, patient demographics, clinical presentation, interventions, and prehospital outcome data. Standard and state-level data collection have influenced participating states year to year, and thus data acquisition. We queried the years 2017-2020 and identified records pertaining to fire dispatches and transports of patients with inhalation injuries. This study was designed in accordance with the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines for observational studies.

Covariates
Relevant covariates extracted were age, sex, race, scene type, and incident disposition.

Outcome measures
The number of fire dispatches and suspected smoke inhalations were used along with the total number of EMS activations, to calculate the raw annual incidence per 100,000 activations. The change in incidence was evaluated using the previous year's data. Poisson regression was then used to estimate the incidence rate ratio (IRR) for each year using 2017 as a baseline.

Statistical Analysis
Counts and proportions were used to summarize categorical variables and non-normally distributed continuous variables were reported as medians with interquartile ranges (IQR). Pearson's Chi squared test was used to compare each year's incidence with the previous year's data. Poisson regression was A c c e p t e d M a n u s c r i p t used to evaluate whether there was an increase in the IRR of fire dispatches and suspected smoke inhalations during the pandemic. Model validation was then performed using deviance and Pearson goodness-of-fit tests. There was no missing data on outcomes of interest; a complete case analysis was undertaken. The threshold for statistical significance was set at p<0.05. Statistical analyses were performed using StataMP release 17 (College Station, TX).

RESULTS
We identified 397,258 fire dispatches out of 107,751,438 EMS dispatches over the study period. In 2020, there was an 8.4% increase in the incidence relative to 2019 (p<0.001) and a 13.4% increase over the average of the previous 3 years (p<0.001). The incidence rate ratio of fire dispatches was also increased in 2020 (1.271; 95% CI: 1.254-1.288; p<0.001; Table 2).
There were 11,732 patients with suspected smoke inhalations during the study period.
Overall, the median age was 44 (IQR: 28-60), 54% were male, and 50% were of unknown race, 74% were injured at home, and 58% were transported to the hospital (Table 1). There was a relative increase of 1.06 per 10,000 from 2017-2020. In 2020, there was a 12% increase in the incidence relative to 2019 (p<0.001) and a 15% increase relative to the average of the previous 3 years (p<0.001). The incidence rate ratio also increased significantly (1.152; 95% CI: 1.070-1.241; p<0.001; Table 3).

DISCUSSION
The scope of the impact of COVID-19 on the United States healthcare system extends beyond resource utilization secondary to the disease itself; it also encompasses downstream effects of lockdown-related lifestyle modifications. Large scale changes in living patterns have led to changes in burn-related injury incidence and severity. Our study found an increase in the incidence and IRR of inhalation injuries, in-hospital mortality, as well as higher total body surface area of full-thickness burns during the lockdown period in Tokyo, Japan [7]. Our results also add context to the results of the 2020 NFPA report. Residential structure fires rose 5% from 2019 to 2020, while non-residential fires fell by eight percent and the rate of deaths per 1,000 home fires was 7.2, compared to 7.1 per 1,000 fires in 1980 [6].
The reasons for this observed increase are undoubtedly multifactorial. The NFPA report suggests changes in home-cooking increased the risk of house fires. Consumer reporting has suggested that up to 55% of Americans were eating at home during the pandemic compared to prelockdown [11]. As more Americans resorted to food preparation at home, it is understandable that this could have contributed to the increase in house fires. Further research will serve to delineate the causation of increased incidence of suspected smoke inhalation. We presume a change in fire prevention efforts also played a role. Many community fire departments across the country engage in outreach events such as home visits for smoke alarm inspection and large-scale community fire prevention events. As social distancing policies dictated events such as these could not be held, it is expected that the lack of fire prevention outreach coupled with more time spent at home engaging in higher risk activities such as cooking only exacerbated the problem. An important future area of research will be to further classify suspected smoke inhalation injuries by region, socioeconomic status, race, and other descriptive demographic data. This will allow better tailoring of community fire prevention measures. Finally, it is unknown whether burn centers in the United States suffered as a whole from the staffing shortages commonly seen during the pandemic. However, some anecdotal literature suggests that operative volume of burn patients increased during lock down periods and as a result the authors stress a need to maintain adequate staffing levels in burn units [12].
A c c e p t e d M a n u s c r i p t This study has the inherent limitations of a retrospective review. The NEMSIS database, though well cited in the literature, has several important limitations to consider. Principally, it is a convenience dataset meaning that not all EMS activations in the country are captured every year as data collection is highly dependent on cooperation from state health agencies. For example, in 2017 NEMSIS implemented its V3 data standard and as a result, not all states transitioned and were hence not represented. Another limitation is the method in which injury information is collected. Patient care reports in the prehospital setting are usually limited to 2 "provider impressions" which are collected by the data standard. Therefore, some inhalation injuries may not have been captured by the dataset if more severe injuries or conditions were recorded instead, superseding and functionally negating the diagnosis of inhalational injury. Finally, in using a national prehospital database, we did not capture suspected smoke inhalations that may have been transported by personal vehicle to an emergency department. It's also important to note that suspected smoke inhalation does not equal an inhalation injury. It is inherently difficult to diagnose an inhalational injury using clinical findings alone. Singed nasal hair for example is the most outward sign and realistically the one of the only signs of IH that can be recognized in the field. However, clinical findings such as these have been shown to have poor discriminatory capability in inhalation injury diagnosis when compared to gold standard bronchoscopy [13]. It is likely that since suspected smoke inhalation is not a true inhalation injury that the estimated incidence is likely an overestimate of true inhalation injury.
In summary, there was an increase in the suspected smoke inhalation exposures and fire dispatches seen in the prehospital setting during the first year of the COVID-19 pandemic compared to previous years. Though this effect is likely multifactorial, various public health and human behavioral factors associated with the pandemic are likely contributory. Further studies should also examine hospital level data to elucidate better the changes in morbidity and mortality at the national level. Our results should serve to inform public health agencies, fire departments, burn centers, and the general public to improve delivery of fire prevention efforts in the event future pandemics to mitigate these trends.
A c c e p t e d M a n u s c r i p t