ORIGINAL ARTICLE
Explaining the Paradoxical Age-Based Racial Disparities in Survival After Trauma The Role of the Treating Facility Caitlin W. Hicks, MD, MS,∗ Zain G. Hashmi, MBBS,∗ Xuan Hui, MD,∗ Catherine Velopulos, MD,∗ David T. Efron, MD,∗ Eric B. Schneider, PhD,∗ Lisa Cooper, MD, MPH,‡ Elliott R. Haut, MD,∗ Edward E. Cornwell III, MD,† and Adil H. Haider, MD, MPH, FACS∗ ‡ Objective: The objective of our study was to determine if differences in outcomes at treating facilities can help explain these age-based racial disparities in survival after trauma. Background: It has been previously demonstrated that racial disparities in survival after trauma are dependent on age. For patients younger than 65 years, blacks had an increased odds of mortality compared with whites, but among patients 65 years or older the opposite association was found. Methods: Data on white and black trauma patients were extracted from the Nationwide Inpatient Sample (2003–2009) using International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis codes. Standardized observed-to-expected mortality ratios were calculated for individual treating facilities, adjusting for age, sex, insurance status, mechanism of injury, overall injury severity, head injury severity, and comorbid conditions. Observed-to-expected ratios were used to benchmark facilities as high-, average-, or low-performing facilities. Proportions and survival outcomes of younger (range, 16–64 years) and older (≥65 years) patients admitted within each performance stratum were compared. Results: A total of 934,476 patients from 1137 facilities (8.3% highperforming, 85% average-performing, and 6.7% low-performing) were analyzed. Younger black patients had a higher adjusted odds of mortality compared with younger white patients [odds ratio, 1.19; 95% confidence interval, 1.11–1.27], whereas older black patients had a lower odds of mortality compared with older white patients [odds ratio, 0.81; 95% confidence interval, 0.74–88]. A significantly greater proportion of younger black patients were treated at low-performing facilities compared with both younger white
From the ∗ Department of Surgery, Center for Surgical Trials and Outcomes Research, Johns Hopkins School of Medicine, Baltimore, MD; †Department of Surgery, Outcomes Research Center, Howard University College of Medicine, Washington, DC; and ‡Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD. Disclosure: Supported by National Institutes of Health/NIGMS K23GM093112-01 and American College of Surgeons C. James Carrico Fellowship for the study of Trauma and Critical Care (A.H.H.). Dr Haut is the primary investigator of the National Institutes of Health Mentored Clinician Scientist Development Award K08 1K08HS017952-01 from the Agency for Healthcare Research and Quality entitled “Does Screening Variability Make DVT an Unreliable Quality Measure of Trauma Care?” Dr Haut receives royalties from Lippincott, Williams, & Wilkins for a book he coauthored (Avoiding Common ICU Errors). He has received honoraria for various speaking engagements regarding clinical and quality and safety topics and has given expert witness testimony in various medical malpractice cases. The data herein was presented as an Oral Presentation at the American College of Surgeons 2013 Clinical Congress in Washington, DC 10/2013 (reference no. SF2013-38680). The remaining authors have no conflicts of interest to disclose. Reprints: Adil H. Haider, MD, MPH, FACS, Department of Surgery, Center for Surgical Trials and Outcomes Research, Johns Hopkins School of Medicine, 1800 Orleans St, Zayed 6107, Baltimore, MD 21287. E-mail:
[email protected]. C 2014 by Lippincott Williams & Wilkins Copyright ISSN: 0003-4932/14/00000-0001 DOI: 10.1097/SLA.0000000000000809
Annals of Surgery r Volume 00, Number 00, 2014
patients and older black patients (49.6% vs 42.2% and 38.7%, respectively; P < 0.05). Conclusions: Nearly half of all young black trauma patients are treated at low-performing facilities. However, facility-based differences do not seem to explain the paradoxical age-based racial disparities after trauma observed in the older population. Keywords: elderly, hospital performance, outcomes, racial disparities, trauma (Ann Surg 2014;00:1–5)
T
he concept of racial disparities after trauma is not new. In a metaanalysis of studies assessing the effects of race on posttraumatic outcomes, black patients were found to be 19% more likely to die after trauma compared with white patients.1 More recently, analysis of data from the Nationwide Inpatient Sample (NIS) suggested that these racial disparities in survival after trauma may be dependent on age.2 In patients younger than 65 years, we demonstrated that black patients had a significantly higher odds of mortality compared with white patients, consistent with previous studies. However, for patients older than 65 years the opposite was true; after correcting for patient comorbidities and other covariates, older black patients actually had a lower odds of death than older white patients. The cause of this paradoxical difference in trauma outcomes is currently unclear. There are some data to suggest hospital-based effects may play a role in race-based disparities after trauma within the younger population. In a nationwide analysis of more than 400 hospitals, the odds of death for patients treated at hospitals with more than 50% minority trauma patients was worse than those treated at hospitals with predominantly white trauma patients.3 In addition, minority trauma patients have been shown to cluster at trauma centers with worse-than-expected mortality rates, which could explain the worse outcomes observed in this population.4 However, the relationship between trauma center performance and age-based racial disparities in trauma mortality is not currently known. In this study, we aim to determine if differences in outcomes at treating facilities can help explain the age-dependent racial differences after trauma that have been previously described. By better defining the risk factors for poor outcomes in an older population, we may identify potentially preventable biases that can be addressed in the current health care system to improve outcomes and reduce racial disparities overall.
METHODS Institutional review board approval was obtained for this study. All white and black adult (16 years or older) trauma patients with blunt and/or penetrating injuries in the NIS for the years 2003–2011 were included. Patients were identified for inclusion using International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis codes (800 through 904 inclusive). Patients from states that www.annalsofsurgery.com | 1
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 00, Number 00, 2014
Hicks et al
report more than 40% data based on race (ie, NC, NV, OR, KY, IL, GA, NE, WA, MT, WV, OH, and MN), hospitals reporting more than 20% missing data on variables of interest, and hospitals with less than 100 eligible patients or that reported no deaths were excluded (Fig. 1). Patients who were transferred to/from another acute care facility, or who had uncertain discharge dispositions, or whose information on important anatomic injury severity was missing were also excluded. Demographic, injury, treating facility, and mortality data were collected for each of the eligible patients. Mortality was defined as in-hospital mortality. Measures of trauma severity, including the injury severity score and mortality prediction model score, were generated for each patient using the International Classification of Diseases Programs for Injury Categorization program5 as previously described.2 A publicly available STATA module was to generate the Charlson Comorbidity Index (CCI)6 from diagnosis codes specific to each patient within the dataset to allow for risk adjustment based on patients’ comorbid conditions. Primary outcome measures included the risk-adjusted odds of mortality and the proportion of patients admitted within each treating facility performance stratum.
Statistical Methods Hospital performance was ascertained using observed-toexpected (O/E) mortality ratios. Individual patient probability of mortality was estimated using a multivariable logistic regression model adjusting for age, sex, race, insurance status, type of injury (blunt vs penetrating), intent of injury, injury severity (trauma mortality
Total number of hospitals (trauma patients) in the NIS 2003-2011 4034 (2,775,830) Excluding states with >40% missing race data 3053 (2,190,855) Excluded hospitals reporting >20% missing data on variables of interest 1637 (1,476,716) Included patients 16 years of age 1636 (1,376,293) Included white or black patients 1627 (1,146!394) Included patients with blunt and penetrating injuries 1627 (1,003,891) Excluded patients with uncertain discharge and/or missing data*" 1622 (955!535) Excluded hospitals with
