Development of a Machine Learning Model to Identify Colorectal Cancer Stage in Medicare Claims

PURPOSE

Staging information is essential for colorectal cancer research. Medicare claims are an important source of population-level data but currently lack oncologic stage. We aimed to develop a claims-based model to identify stage at diagnosis in patients with colorectal cancer.

METHODS

We included patients age 66 years or older with colorectal cancer in the SEER-Medicare registry. Using patients diagnosed from 2014 to 2016, we developed models (multinomial logistic regression, elastic net regression, and random forest) to classify patients into stage I-II, III, or IV on the basis of demographics, diagnoses, and treatment utilization identified in Medicare claims. Models developed in a training cohort (2014-2016) were applied to a testing cohort (2017), and performance was evaluated using cancer stage listed in the SEER registry as the reference standard.

RESULTS

The cohort of patients with 30,543 colorectal cancer included 14,935 (48.9%) patients with stage I-II, 9,203 (30.1%) with stage III, and 6,405 (21%) with stage IV disease. A claims-based model using elastic net regression had a scaled Brier score (SBS) of 0.45 (95% CI, 0.43 to 0.46). Performance was strongest for classifying stage IV (SBS, 0.62; 95% CI, 0.59 to 0.64; sensitivity, 93%; 95% CI, 91 to 94) followed by stage I-II (SBS, 0.45; 95% CI, 0.44 to 0.47; sensitivity, 86%; 95% CI, 85 to 76) and stage III (SBS, 0.32; 95% CI, 0.30 to 0.33; sensitivity, 62%; 95% CI, 61 to 64).

CONCLUSION

Machine learning models effectively classified colorectal cancer stage using Medicare claims. These models extend the ability of claims-based research to risk-adjust and stratify by stage.

Early VTE Prophylaxis in Severe Traumatic Brain Injury: A Propensity Score Weighted EAST Multi-Center Study

BACKGROUND

Patients with TBI are at high risk of venous thromboembolism events (VTE). We hypothesized that early chemical VTE prophylaxis initiation (≤24 hours of a stable head CT) in severe TBI would reduce VTE without increasing risk of intracranial hemorrhage expansion (ICHE).

METHODS

A retrospective review of adult patients ≥18 years of age with isolated severe TBI (AIS ≥ 3) who were admitted to 24 level 1 and level 2 trauma centers from January 1, 2014 to December 31 2020 was conducted. Patients were divided into those who did not receive any VTE prophylaxis (NO VTEP), who received VTE prophylaxis ≤24 hours after stable head CT (VTEP ≤24) and who received VTE prophylaxis >24 hours after stable head CT (VTEP>24). Primary outcomes were VTE and ICHE. Covariate balancing propensity score weighting was utilized to balance demographic & clinical characteristics across three groups. Weighted univariate logistic regression models were estimated for VTE & ICHE with patient group as predictor of interest.

RESULTS

Of 3,936 patients, 1,784 met inclusion criteria. Incidences of VTE was significantly higher in the VTEP>24 group, with higher incidences of DVT in the group. Higher incidences of ICHE were observed in the VTEP≤24 and VTEP>24 groups. After propensity score weighting, there was a higher risk of VTE in patients in VTEP >24 compared to those in VTEP≤24 ( [OR] = 1.51; [95%CI] = 0.69-3.30; p = 0.307), however was not significant. Although, the No VTEP group had decreased odds of having ICHE compared to VTEP≤24 (OR = 0.75; 95%CI = 0.55-1.02, p = 0.070), the result was not statistically significant.

CONCLUSIONS

In this large multi-center analysis, there were no significant differences in VTE based on timing of initiation of VTE prophylaxis. Patients who never received VTE prophylaxis had decreased odds of ICHE. Further evaluation of VTE prophylaxis in larger randomized studies will be necessary for definitive conclusions.

LEVEL OF EVIDENCE

level III, Therapeutic Care Management.

Youth Firearm Injury: A Review for Pediatric Critical Care Clinicians

Firearms are now the leading cause of death among youth in the United States, with rates of homicide and suicide rising even more steeply during the SARS-CoV-2 pandemic. These injuries and deaths have wide-ranging consequences for the physical and emotional health of youth and families. While pediatric critical care clinicians must treat the injured survivors, they can also play a role in prevention by understanding the risks and consequences of firearm injuries; taking a trauma-informed approach to the care of injured youth; counseling patients and families on firearm access; and advocating for youth safety policy and programming.

Higher center volume is significantly associated with lower mortality in trauma patients with shock

INTRODUCTION

Injured patients presenting in shock are at high risk of mortality despite numerous efforts to improve resuscitation. Identifying differences in outcomes among centers for this population could yield insights to improve performance. We hypothesized that trauma centers treating higher volumes of patients in shock would have lower risk-adjusted mortality.

METHODS

We queried the Pennsylvania Trauma Outcomes Study from 2016 to 2018 for injured patients ≥16 years of age at Level I&II trauma centers who had an initial systolic blood pressure (SBP) of <90 mmHg. We excluded patients with critical head injury (abbreviated injury score [AIS] head ≥5) and patients coming from centers with a shock patient volume of ≤10 for the study period. The primary exposure was tertile of center-level shock patient volume (low, medium, or high volume). We compared risk-adjusted mortality by tertile of volume using multivariable Cox proportional hazards model incorporating age, injury severity, mechanism, and physiology.

RESULTS

Of 1,805 included patients at 29 centers, 915 (50.7%) died. The median annual shock trauma patient volume was 9 patients for low volume centers, medium 19.5, and high 37. Median ISS was higher at high volume compared to low volume centers (22 vs 18, p <0.001). Raw mortality was 54.9% at high volume centers, 46.7% for medium, and 42.9% for low. Time elapsed from arrival to emergency department (ED) to the operating room (OR) was lower at high volume than low volume centers (median 47 vs 78 min) p = 0.003. In adjusted analysis, hazard ratio for high volume centers (referenced to low volume) was 0.76 (95% CI 0.59-0.97, p = 0.030).

CONCLUSION

After adjusting for patient physiology and injury characteristics, center-level volume is significantly associated with mortality. Future studies should seek to identify key practices associated with improved outcomes in high-volume centers. Furthermore, shock patient volume should be considered when new trauma centers are opened.

Law enforcement in the trauma bay: a survey of members of the American Academy for the Surgery of Trauma

Background

Trauma patients frequently come into contact with law enforcement officers (LEOs) during the course of their medical care, but little is known about how LEO presence affects processes of care. We surveyed members of the American Association for the Surgery of Trauma (AAST) to assess their perspectives on frequency, circumstances, and implications of LEO presence in trauma bays nationwide.

Methods

Survey items addressed respondents' experience with the frequency and context of LEO presence and their perspectives on the impact of LEO presence for patients, clinical care, and public safety. Respondent demographics, professional characteristics, and practice setting were collected. The survey was distributed electronically to AAST members in September and October of 2020. Responses were compared by participant age, gender, race, ethnicity, urban versus rural location using χ2 tests.

Results

Of 234 respondents, 189 (80.7%) were attending surgeons, 169 (72.2%) identified as white, and 144 (61.5%) as male. 187 respondents (79.9%) observed LEO presence at least weekly. Respondents found LEO presence was most helpful for public safety, followed by clinical care, and then for patients. Older respondents rated LEO presence as helpful more often than younger respondents regarding the impact on patients, clinical care, and public safety (p<0.001 across all domains). When determining LEO access, respondents assessed severity of the patient's condition, the safety of emergency department staff, the safety of LEOs, and a patient's potential role as a threat to public safety.

Conclusions

Respondents described a wide range of perspectives on the impact and consequence of LEO in the trauma bay, with little policy to guide interactions. The overlap of law enforcement and healthcare in the trauma bay deserves attention from institutional and professional policymakers to preserve patient safety and autonomy and patient-centered care.

Level of evidence

IV, survey study.

Preventing the most common firearm deaths: Modifiable factors related to firearm suicide

BACKGROUND

More than 20,000 firearm suicides occur every year in America. Firearm restrictive legislation, firearm access, demographics, behavior, access to care, and socioeconomic metrics have been correlated to firearm suicide rates. Research to date has largely evaluated these contributors singularly. We aimed to evaluate them together as they exist in society. We hypothesized that state firearm laws would be associated with reduced firearm suicide rates.

METHODS

We acquired the 2013 to 2016 data for firearm suicide rates from The Centers for Disease Control Wide-ranging Online Data for Epidemiologic Research. Firearm laws were obtained from the State Firearms Law Database. Depression rates and access to care were obtained from the Behavioral Risk Factor Surveillance System and Occupational Employment and Wage Statistics program. Population demographics, poverty, and access to social support were obtained from the American Community Survey. Firearm access estimates were retrieved from the National Instant Criminal Background Check System. We used a univariate panel linear regression with fixed effect for state and firearm suicide rates as the outcome. We created a final multivariable model to determine the adjusted associations of these factors with firearm suicide rates.

RESULTS

In univariate analysis, firearm access, heavy drinking behavior, demographics, and access to care correlated to increased firearm suicide rates. The state proportion identifying as white and the proportion of those in poverty receiving food benefits correlated to decreased firearm suicide rates. In multivariable regression, only heavy drinking (β, 0.290; 95% confidence interval, 0.092-0.481; P = .004) correlated to firearm suicides rates increases.

CONCLUSIONS

During our study, few firearm laws changed. Heavy drinking behavior association with firearm suicide rates suggests an opportunity for interventions exists in the health care setting.

Injury characteristics and circumstances of firearm trauma: Assessing suicide survivors and decedents

INTRODUCTION

Despite representing fewer than 5% of suicide attempts, firearms account for over half of deaths. Yet there is little clinical information regarding firearm attempts, particularly survivors. We assessed clinical factors differentiating firearm suicide survivors from decedents, firearm attempters from other methods, and firearm attempters from similarly injured trauma patients.

METHODS

We used clinical data from the National Trauma Data Bank (2017) to assess firearm suicide attempts using cross-sectional and case-control designs. We used logistic and multinomial regression to compare groups and assess firearm type and discharge destination.

RESULTS

Older age, being uninsured, and injury location were associated with increased mortality among firearm attempters. Older age, White race, male sex, and being uninsured were associated with firearm attempts. Major psychiatric disorders were associated with firearm attempts and using a rifle or shotgun. Major psychiatric disorders, female sex, and smoking were associated with psychiatric discharge. Black and other race were associated with law enforcement discharge, and Black race was associated with lower odds of psychiatric discharge. Uninsured patients had lower odds of discharge to long-term care, psychiatric, or rehabilitation facilities.

CONCLUSIONS

This study identifies factors associated with firearm suicide and includes indicators of disparities in health services for patients at high risk of suicide death.

The variable role of damage control laparotomy over 19 years of trauma care in Pennsylvania

BACKGROUND

Damage control laparotomy emphasizes physiologic stabilization of critically injured patients and allows staged surgical management. However, there is little consensus on the optimal criteria for damage control laparotomy. We examined variability between centers and over time in Pennsylvania.

METHODS

We analyzed the Pennsylvania Trauma Outcomes Study data between 2000 and 2018, excluding centers performing <10 laparotomies/year. Laparotomy was defined using International Classification of Diseases codes, and damage control laparotomy was defined by a code for "reopening of recent laparotomy" or a return to the operating room >4 hours from index laparotomy that was not unplanned. We examined trends over time and by center. Multivariable logistic regression models were developed to predict both damage control laparotomy and mortality, generate observed:expected ratios, and identify outliers for each. We compared risk-adjusted mortality rates to center-level damage control laparotomy rates.

RESULTS

In total, 18,896 laparotomies from 22 centers were analyzed; 3,549 damage control laparotomies were performed (18.8% of all laparotomies). The use of damage control laparotomy in Pennsylvania varied from 13.9% to 22.8% over time. There was wide variation in center-level use of damage control laparotomy, from 11.1% to 29.4%, despite adjustment. Factors associated with damage control laparotomy included injury severity and admission vital signs. Center identity improved the model as demonstrated by likelihood ratio test (P < .001), suggesting differences in center-level practices. There was minimal correlation between center-level damage control laparotomy use and mortality.

CONCLUSION

There is wide center-level variation in the use of damage control laparotomy among centers, despite adjustment for patient factors. Damage control laparotomy is both resource intensive and highly morbid; regional resources should be allocated to address this substantial practice variation to optimize damage control laparotomy use.

Rates of Surgical Consultations After Emergency Department Admission in Black and White Medicare Patients

Importance

A surgical consultation is a critical first step in the care of patients with emergency general surgery conditions. It is unknown if Black Medicare patients and White Medicare patients receive surgical consultations at similar rates when they are admitted from the emergency department.

Objective

To determine whether Black Medicare patients have similar rates of surgical consultations when compared with White Medicare patients after being admitted from the emergency department with an emergency general surgery condition.

Design, Setting, and Participants

This was a retrospective cohort study that took place at US hospitals with an emergency department and used a computational generalization of inverse propensity score weight to create patient populations with similar covariate distributions. Participants were Medicare patients age 65.5 years or older admitted from the emergency department for an emergency general surgery condition between July 1, 2015, and June 30, 2018. The analysis was performed during February 2022. Patients were classified into 1 of 5 emergency general surgery condition categories based on principal diagnosis codes: colorectal, general abdominal, hepatopancreatobiliary, intestinal obstruction, and upper gastrointestinal.

Exposures

Black vs White race.

Main Outcomes and Measures

Receipt of a surgical consultation after admission from the emergency department with an emergency general surgery condition.

Results

A total of 1 686 940 patients were included in the study. Of those included, 214 788 patients were Black (12.7%) and 1 472 152 patients were White (87.3%). After standardizing for medical and diagnostic imaging covariates, Black patients had 14% lower odds of receiving a surgical consultation (odds ratio [OR], 0.86; 95% CI, 0.85-0.87) with a risk difference of -3.17 (95% CI, -3.41 to -2.92). After standardizing for socioeconomic covariates, Black patients remained at an 11% lower odds of receiving a surgical consultation compared with similar White patients (OR, 0.89; 95% CI, 0.88-0.90) with a risk difference of -2.49 (95% CI, -2.75 to -2.23). Additionally, when restricting the analysis to Black patients and White patients who were treated in the same hospitals, Black patients had 8% lower odds of receiving a surgical consultation when compared with White patients (OR, 0.92; 95% CI, 0.90-0.93) with a risk difference of -1.82 (95% CI, -2.18 to -1.46).

Conclusions and Relevance

In this study, Black Medicare patients had lower odds of receiving a surgical consultation after being admitted from the emergency department with an emergency general surgery condition when compared with similar White Medicare patients. These disparities in consultation rates cannot be fully attributed to medical comorbidities, insurance status, socioeconomic factors, or individual hospital-level effects.

Analyzing Impact of Multimorbidity on Long-Term Outcomes after Emergency General Surgery: A Retrospective Observational Cohort Study

BACKGROUND

Little is known about the impact of multimorbidity on long-term outcomes for older emergency general surgery patients.

STUDY DESIGN

Medicare beneficiaries, age 65 and older, who underwent operative management of an emergency general surgery condition were identified using Centers for Medicare & Medicaid claims data. Patients were classified as multimorbid based on the presence of a Qualifying Comorbidity Set (a specific combination of comorbid conditions known to be associated with increased risk of in-hospital mortality in the general surgery setting) and compared with those without multimorbidity. Risk-adjusted outcomes through 180 days after discharge from index hospitalization were calculated using linear and logistic regressions.

RESULTS

Of 174,891 included patients, 45.5% were identified as multimorbid. Multimorbid patients had higher rates of mortality during index hospitalization (5.9% vs 0.7%, odds ratio [OR] 3.05, p < 0.001) and through 6 months (17.1% vs 3.4%, OR 2.33, p < 0.001) after discharge. Multimorbid patients experienced higher rates of readmission at 1 month (22.9% vs 11.4%, OR 1.48, p < 0.001) and 6 months (38.2% vs 21.2%, OR 1.48, p < 0.001) after discharge, lower rates of discharge to home (42.5% vs 74.2%, OR 0.52, p < 0.001), higher rates of discharge to rehabilitation/nursing facility (28.3% vs 11.3%, OR 1.62, p < 0.001), greater than double the use of home oxygen, walker, wheelchair, bedside commode, and hospital bed (p < 0.001), longer length of index hospitalization (1.33 additional in-patient days, p < 0.001), and higher costs through 6 months ($5,162 additional, p < 0.001).

CONCLUSIONS

Older, multimorbid patients experience worse outcomes, including survival and independent function, after emergency general surgery than nonmultimorbid patients through 6 months after discharge from index hospitalization. This information is important for setting recovery expectations for high-risk patients to improve shared decision-making.

The Epidemiology of Firearm Injuries in the US: The Need for Comprehensive, Real-time, Actionable Data

This Viewpoint summarizes current data on firearm injuries in the US, discusses the limitations of available data sources, and proposes measures for a comprehensive system to track firearm injury and death

Multimorbidity Confers Greater Risk for Older Patients in Emergency General Surgery Than the Presence of Multiple Comorbidities: A Retrospective Observational Study

BACKGROUND

Little is known about the impact of multimorbidity on outcomes for older emergency general surgery patients.

OBJECTIVE

The aim was to understand whether having multiple comorbidities confers the same amount of risk as specific combinations of comorbidities (multimorbidity) for a patient undergoing emergency general surgery.

RESEARCH DESIGN

Retrospective observational study using state discharge data.

SUBJECTS

Medicare beneficiaries who underwent an operation for an emergency general surgery condition in New York, Florida, or Pennsylvania (2012-2013).

MEASURES

Patients were classified as multimorbid using Qualifying Comorbidity Sets (QCSs). Outcomes included in-hospital mortality, hospital length of stay and discharge status.

RESULTS

Of 312,160 patients, a large minority (37.4%) were multimorbid. Non-QCS patients did not have a specific combination of comorbidities to satisfy a QCS, but 64.1% of these patients had 3+ comorbid conditions. Multimorbidity was associated with increased in-hospital mortality (10.5% vs. 3.9%, P <0.001), decreased rates of discharge to home (16.2% vs. 37.1%, P <0.001), and longer length of stay (10.4 d±13.5 vs. 6.7 d±9.3, P <0.001) when compared with non-QCS patients. Risks varied between individual QCSs.

CONCLUSIONS

Multimorbidity, defined by satisfying a specific QCS, is strongly associated with poor outcomes for older patients requiring emergency general surgery in the United States. Variation in risk of in-hospital mortality, discharge status, and length of stay between individual QCSs suggests that multimorbidity does not carry the same prognostic weight as having multiple comorbidities-the specifics of which are important in setting expectations for individual, complex patients.

Demographic and Regional Factors Associated With Reporting Homicides of Transgender People in the United States

INTRODUCTION

The American Medical Association recently declared homicides of transgender individuals an epidemic. However, transgender homicide victims are often classified as nontransgender. Our objective was to describe existing data and coding of trans (i.e., transgender) victims and to examine the risk factors for homicides of trans people relative to nontrans people across the United States.

METHODS

A retrospective review of the Centers for Disease Control and Prevention's National Violent Death Reporting System for the years 2003-2018 identified victims defined as transgender either through the "transgender" variable or narrative reports. Fisher's exact tests and logistic regression models were run to compare the demographics of trans victims to those not identified as trans.

RESULTS

Of the 147 transgender victims identified, 14.4% were incorrectly coded as nontrans despite clear indication of trans status in the narrative description, and 6% were coded as hate crimes. Relative to nontrans victims, trans victims were more frequently Black (54.4% versus 40.7%, P = 0.001), had a mental health condition (26.5% versus 11.3%, P < 0.001), or reported being a sex worker (9.5% versus 0.2%, P < 0.001). There were disproportionately few homicides of transgender people in the South (13.6% of trans victims versus 29.1% of nontrans victims, P < 0.001). Conversely, the West and Midwest accounted for a higher-than-expected proportion of trans victims relative to nontrans victims (23.1% of trans victims versus 16.2% of nontrans victims, P = 0.03; 24.5% of trans victims versus 16.8% of nontrans victims, P = 0.02, respectively).

CONCLUSIONS

Though the murder of transgender individuals is a known public health crisis, inconsistencies still exist in the assessment and reporting of transgender status. Further, these individuals were more likely to have multiple distinct vulnerabilities. These findings provide important information for injury and violence prevention researchers to improve reporting of transgender status in the medical record and local trauma registries.

County-Level Variation in Changes in Firearm Mortality Rates Across the US, 1989 to 1993 vs 2015 to 2019

IMPORTANCE

Firearm violence remains a critical public health challenge, disproportionately impacting some US regions. County-level variation may hold key insights into how firearm mortality rates vary across the US.

OBJECTIVE

To model county-level changes in firearm mortality rates (total, homicide, and suicide) from 1989 to 1993 vs 2015 to 2019 and identify and characterize hot spots showing unexpected changes over time.

DESIGN, SETTING, AND PARTICIPANTS

This is a cross-sectional study with 2 time points using a novel small area estimation method to analyze restricted access mortality microdata by cause of death and US county. The analysis included 3111 US counties from 49 states and the District of Columbia from January 1, 1989, to December 31, 2019. Bayesian spatial models were fitted to map geographical variation in changes in age-standardized firearm mortality rates (per 100 000 person-years) from 1989 to 1993 vs 2015 to 2019. County outliers (or hot spots) were defined as having observed rates that fell outside the 95% credible intervals of their expected posterior predictive distribution. These counties were characterized using visualization and descriptive statistics of their characteristics. Data were analyzed from June to December 2021.

EXPOSURES

County of residence.

MAIN OUTCOMES AND MEASURES

Five-year age-standardized mortality rates by US county, age, and cause of death for 1989 to 1993 and 2015 to 2019.

RESULTS

Between 1989 and 2019, 1 036 518 firearm deaths were recorded in counties across the US. Suicide was the most common cause of firearm mortality (589 285 deaths) followed by homicide (412 231 deaths). Age-standardized rates (deaths per 100 000 individuals) for firearm deaths and suicides increased from 1989 to 1993 vs 2015 to 2019 (mean [SD] change, 0.16 [8.78] for firearm deaths and 1.21 [6.91] for suicides), while firearm homicides decreased (mean [SD] change, -0.39 [3.96]). However, these national trends were not homogeneous across counties and often varied by geographical region. The West and Midwest showed the most pronounced increases in firearm suicide rates, whereas the Southeast showed localized increases in firearm homicide rates, despite the national decreasing trend. Critical hot spots were identified in urban counties of Alabama, and firearm homicide rates (per 100 000) in Baltimore City, Maryland, almost doubled from 29.71 to 47.43, and by 2015 to 2019 it accounted for 66.7% of all firearm homicide in Maryland. By contrast, District of Columbia showed promising improvements over time, decreasing from 56.5 firearm homicides per 100 000 in 1989 to 1993 to 14.45 in 2015 to 2019.

CONCLUSIONS AND RELEVANCE

There was substantial variation in rates and changes in firearm deaths among US counties. Geographical hot spots may be useful to inform targeted prevention efforts and local policy responses.

Compounding Trauma: the Intersections of Racism, Law Enforcement, and Injury

Purpose of Review

Traumatic injury sits at the nexus of law enforcement and structural racism. This narrative review aims to explore the major impacts of law enforcement on health, its intersections with US structural racism, and their joint impacts on traumatic injury and injury care.

Recent Findings

Many of the same forces of systemic disadvantage that put Black people, other people of color, and other marginalized groups at risk for violent injury also expose these same individuals and communities to intensive policing. Recent evidence speaks to the broad impact of police exposure and police violence on individual and community physical and mental health. Moreover, injured patients who are exposed to law enforcement during their care are at risk for erosion of trust in and relationships with their healthcare providers. To optimize the role of law enforcement agencies in injury prevention, collaboration across sectors and with communities is essential.

Summary

A broad approach to the prevention of injury and violence must incorporate an understanding of the intersecting impacts of law enforcement and structural racism on health and traumatic injury. Clinicians who seek to provide trauma-informed injury care should incorporate an understanding of the role of law enforcement in individual and community health.

The changing epidemiology of interpersonal firearm violence during the COVID-19 pandemic in Philadelphia, PA

Recent increases in firearm violence in U.S. cities are well-documented, however dynamic changes in the people, places and intensity of this public health threat during the COVID-19 pandemic are relatively unexplored. This descriptive epidemiologic study spanning from January 1, 2015 - March 31, 2021 utilizes the Philadelphia Police Department's registry of shooting victims, a database which includes all individuals shot and/or killed due to interpersonal firearm violence in the city of Philadelphia. We compared victim and event characteristics prior to the pandemic with those following implementation of pandemic containment measures. In this study, containment began on March 16, 2020, when non-essential businesses were ordered to close in Philadelphia. There were 331 (SE = 13.9) individuals shot/quarter pre-containment vs. 545 (SE = 66.4) individuals shot/quarter post-containment (p = 0.031). Post-containment, the proportion of women shot increased by 39% (95% CI: 1.21, 1.59), and the proportion of children shot increased by 17% (95% CI: 1.00, 1.35). Black women and children were more likely to be shot post-containment (RR 1.11, 95% CI: 1.02, 1.20 and RR 1.08, 95% CI: 1.03, 1.14, respectively). The proportion of mass shootings (≥4 individuals shot within 100 m within 1 h) increased by 53% post-containment (95% CI: 1.25, 1.88). Geographic analysis revealed relative increases in all shootings and mass shootings in specific city locations post-containment. The observed changes in firearm injury epidemiology following COVID-19 containment in Philadelphia demonstrate an intensification in firearm violence, which is increasingly impacting people who are likely made more vulnerable by existing social and structural disadvantage. These findings support existing knowledge about structural causes of interpersonal firearm violence and suggest structural solutions are required to address this public health threat.

Disparities in Insurance Status are Associated With Outcomes But Not Timing of Trauma Care

INTRODUCTION

Patient factors influence outcomes after injury. Delays in care have a crucial impact. We investigated the associations between patient characteristics and timing of transfer from the emergency department to definitive care.

METHODS

This was a review of adult trauma patients treated between January 1, 2016, and December 31, 2018. Bivariate analyses were used to build Cox proportional hazards models. We built separate logistic and negative binomial regression models for secondary outcomes using mixed-step selection to minimize the Akaike information criterion c.

RESULTS

A total of 1219 patients were included; 68.5% were male, 56.8% White, 11.2% Black, and 7.8% Asian/Pacific Islander. The average age was 51 ± 21 y. Overall, 13.7% of patients were uninsured. The average length of stay was 5 d and mortality was 5.9%. Shorter transfer time out of the emergency department was associated with higher tier of activation (relative risk [RR] 1.39, 95% confidence interval [CI] 1.09-1.77; P = 0.0074), Injury Severity Score between 16 and 24 points (RR 1.57, 95% CI 1.04-2.32; P = 0.0307) or ≥25 (RR 3.85, 95% CI 2.45-5.94; P = 0.0001), and penetrating injury. Longer time to event was associated with Glasgow coma scale score ≥14 points (RR 0.47, 95% CI 0.27-0.85; P = 0.0141). Uninsured patients were less likely to be admitted (odds ratio 0.29, 95% CI 0.17-0.48; P = 0.0001) and more likely to experience shorter length of stay (incidence rate ratio 0.34, 95% CI 0.24-0.51; P = 0.0001).

CONCLUSIONS

Injury characteristics and insurance status were associated with patient outcomes in this retrospective, single-center study. We found no disparity in timing of intrafacility transfer, perhaps indicating that initial management protocols preserve equity.

Increase in Motor Vehicle Crash Severity: An Unforeseen Consequence of COVID-19

INTRODUCTION

The 2019 coronavirus (COVID-19) pandemic led to stay-at-home (SAH) orders in Pennsylvania targeted at reducing viral transmission. Limitations in population mobility under SAH have been associated with decreased motor vehicle collisions (MVC) and related injuries, but the impact of these measures on severity of injury remains unknown. The goal of this study is to measure the incidence, severity, and outcomes of MVC-related injuries associated with SAH in Pennsylvania.

MATERIALS & METHODS

We conducted a retrospective geospatial analysis of MVCs during the early COVID-19 pandemic using a state-wide trauma registry. We compared characteristics of patients with MVC-related injuries admitted to Pennsylvania trauma centers during SAH measures (March 21-July 31, 2020) with those from the corresponding periods in 2018 and 2019. We also compared incidence of MVCs for each zip code tabulation area (ZCTA) in Pennsylvania for the same time periods using geospatial mapping.

RESULTS

Of 15,550 trauma patients treated during the SAH measures, 3486 (22.4%) resulted from MVCs. Compared to preceding years, MVC incidence decreased 10% under SAH measures with no change in mortality rate. However, in ZCTA where MVC incidence decreased, there was a 16% increase in MVC injury severity.

CONCLUSIONS

Stay-at-home orders issued in response to the COVID-19 pandemic in Pennsylvania were associated with significant changes in MVC incidence and severity. Identifying such changes may inform resource allocation decisions during future pandemics or SAH events.

The impact of COVID-19 infection on outcomes after injury in a state trauma system

BACKGROUND

The COVID-19 pandemic reshaped the health care system in 2020. COVID-19 infection has been associated with poor outcomes after orthopedic surgery and elective, general surgery, but the impact of COVID-19 on outcomes after trauma is unknown.

METHODS

We conducted a retrospective cohort study of patients admitted to Pennsylvania trauma centers from March 21 to July 31, 2020. The exposure of interest was COVID-19 (COV+) and the primary outcome was inpatient mortality. Secondary outcomes were length of stay and complications. We compared demographic and injury characteristics between positive, negative, and not-tested patients. We used multivariable regression with coarsened exact matching to estimate the impact of COV+ on outcomes.

RESULTS

Of 15,550 included patients, 8,170 (52.5%) were tested for COVID-19 and 219 (2.7%) were positive (COV+). Compared with COVID-19-negative (COV-) patients, COV+ patients were similar in terms of age and sex, but were less often white (53.5% vs. 74.7%, p < 0.0001), and more often uninsured (10.1 vs. 5.6%, p = 0.002). Injury severity was similar, but firearm injuries accounted for 11.9% of COV+ patients versus 5.1% of COV- patients (p < 0.001). Unadjusted mortality for COV+ was double that of COV- patients (9.1% vs. 4.7%, p < 0.0001) and length of stay was longer (median, 5 vs. 4 days; p < 0.001). Using coarsened exact matching, COV+ patients had an increased risk of death (odds ratio [OR], 6.05; 95% confidence interval [CI], 2.29-15.99), any complication (OR, 1.85; 95% CI, 1.08-3.16), and pulmonary complications (OR, 5.79; 95% CI, 2.02-16.54) compared with COV- patients.

CONCLUSION

Patients with concomitant traumatic injury and COVID-19 infection have elevated risks of morbidity and mortality. Trauma centers must incorporate an understanding of these risks into patient and family counseling and resource allocation during this pandemic.

LEVEL OF EVIDENCE

Level II, Prognostic Study.