M. C. Smith1,2, P. J. Chung1,2, V. Roudnitsky1,2, G. Sugiyama1 1SUNY Downstate Medical Center,Surgery,Brooklyn, NY, USA 2Kings County Hospital Center,Surgery,Brooklyn, NY, USA
Introduction: Suicide is a large public health crisis in the United States, with over 400,000 people treated in emergency departments in 2013 with self-inflicted injuries, and it was responsible for over 41,000 deaths. This makes it the tenth leading cause of death in the country, and the second leading cause of death in the 15 to 35-year-old age group. Using a large, nationwide database, we sought to examine the outcomes of patients who present with self-inflicted gunshot wounds, and to determine risk factors for mortality.
Methods: Data was collected from the Nationwide Inpatient Sample (NIS) from 2005-2012 using ICD 9 E-codes for self-inflicted gunshot wounds (E955, E955.0-E955.4). Patients age <18 years and with missing gender data were excluded. We then calculated the Trauma Mortality Prediction Model (TMPM) scores for each patient to provide an estimated assessment of trauma severity. We also calculated the Elixhauser-Van Walraven score to assess comorbidity status of each patient. We then performed multiple imputation for missing data. Using inpatient mortality as our primary outcome, we then performed multivariable logistic regression to adjust for age, gender, race, insurance status, income status, elective procedure, hospital size, rural vs urban hospital, geographic region, private vs government facility, TMPM score, Elixhauser-Van Walraven score, psychiatric history, and history of depression.
Results: A total of 3,349 patients sustained self-inflicted gunshot wounds. The majority of these patients were male (80.56%) and White (80.53%). The mortality rate was 33.17%. Variables associated with mortality on multivariable logistic regression analysis included age (OR 1.93 [1.61 – 2.32 95% CI], p <0.0001), lack of medical insurance (OR 1.49 [1.12 – 1.98 95% CI], p =0.0185), and TMPM score (OR 6.10 [5.26 – 7.07 95% CI], p <0.0001). Variables associated with decreased risk of mortality included female gender (OR 0.63 [0.47 – 0.84 95% CI], p=0.0014), admission to medium vs large hospital (OR 0.65 [0.50 – 0.86 95% CI], p=0.0057), admission to a rural vs urban teaching hospital (OR 0.27 [0.14 – 0.53 95% CI], p=0.0006), history of psychosis (OR 0.09 [0.06 – 0.13 95% CI], p <0.0001), and history of depression (OR 0.32 [0.25 – 0.42 95% CI], p <0.0001).
Conclusion: In this large, retrospective analysis of the NIS database, we found that among patients with self-inflicted gunshot wounds, age, lack of medical insurance, and increased TMPM score were associated with significantly greater risk of mortality. Factors which decreased mortality were female gender, admission to medium sized hospitals or those in rural areas, and history of psychosis or depression. Further studies should examine these cases on a prospective basis to help mitigate these risks, as well as work towards prevention.