Insurance Status, Not Race, Is a Determinant of Outcomes from Vehicular Injury

Insurance Status, Not Race, Is a Determinant of Outcomes from Vehicular Injury Joseph J Tepas III, MD, FACS, Etienne E Pracht, PhD, Barbara L Orban, PhD, Lewis M Flint, MD, FACS Hypothesizing that outcomes from specific injury mechanisms should not vary by race or socioeconomic status, we analyzed the relationship of race and ethnicity to fatality in motor vehicle crash victims treated during 2008 and 2009. STUDY DESIGN: Logistic regression analysis of pooled administrative data assessed the contribution of patient demographics and injury severity to outcome, defined as mortality during acute hospitalization. Demographic factors included age, sex, race, ethnicity, and insurance. Severe injury was defined using ICD-9 Injury Severity Score (survival probability) p ⬍ 0.85, presence of up to 3 comorbidities, and/or diagnosis of spinal cord injury and/or traumatic brain injury. Mortality was stratified by survival time after trauma center arrival to death within 24 hours or thereafter. Factors contributing to outcomes were tested using chi square analysis of the calculated model estimate. RESULTS: For 8,758 motor vehicle crash victims treated in state-designated trauma centers, age, sex, injury severity, and 2 or more comorbidities consistently predicted survival. Neither race nor ethnicity was associated with increased mortality risk. Being uninsured was related to death within 24 hours (p ⬍ 0.001).The majority of the uninsured who died within 24 hours had an ICD-9 Injury Severity Score p ⱕ 0.5. Mortality risk after 24 hours was driven by traumatic brain injury and comorbidities. CONCLUSIONS: The results of this study indicated that higher immediate mortality of the uninsured is a behavioral and socioeconomic rather than physiologic marker. This higher mortality is driven by increased injury severity that increases cost of care in uninsured survivors. This disparity suggests that risk-taking behavior, especially relating to safety practices and licensing regulations, is an important etiologic factor. Improved outcomes require better public education and enforcement in conjunction with improvements in processes of care. (J Am Coll Surg 2011; 212:722–729. © 2011 by the American College of Surgeons) BACKGROUND:

colleagues,6 in their analysis of outcome disparity, reported a higher mortality from injury in uninsured children and adolescents. The 3 potential causes for this disparity were limitation to access, lower level of health literacy, or different methods of care based on potential reimbursement. Additional data from reports by Crompton and coauthors1 and Maybury and associates7 suggest that the outcome disparities observed in patients categorized by race and insurance status are not confined to specific subsets defined by injury mechanism. We hypothesized that the reasons for this alleged disparity relate to factors independent of access to emergency services and the clinical process of acute injury care. We tested this hypothesis by analyzing adult fatality from motor vehicle crashes treated in a mature state trauma system during 2008 and 2009.

Although trauma systems aspire to equal treatment for all patients, the disease of injury is not so egalitarian. In various forms, injury afflicts virtually every segment of society, with the greatest impact, in terms of death and disability, affecting society’s young and most productive members. In mature trauma systems that are legislatively mandated and monitored for quality, clinical outcomes of specific injury mechanisms, as defined in terms of lost lives and unrealized human potential, should not vary by race or socioeconomic status. Yet recent reports suggest the opposite.1-5 Rosen and Disclosure Information: Nothing to disclose. Presented at Southern Surgical Association 122nd Annual Meeting, Palm Beach, FL, December 2010. Received December 6, 2010; Accepted December 14, 2010. From the Department of Surgery, University of Florida Health Science Center (Tepas) and the College of Public Health (Orban) and Department of Health Policy & Management (Flint), University of South Florida, Jacksonville, FL; and American College of Surgeons, Chicago, IL (Pracht). Correspondence address: Joseph J Tepas III, MD, FACS, Department of Surgery, University of Florida Health Science Center, 655 W 8th St, Jacksonville, FL 32209.

© 2011 by the American College of Surgeons Published by Elsevier Inc.

METHODS The Florida Agency for Health Care Administration is required by statute to collect discharge data from all licensed hospitals in the state. These data include basic demo-

722

ISSN 1072-7515/11/$36.00 doi:10.1016/j.jamcollsurg.2010.12.016

Vol. 212, No. 4, April 2011

Abbreviations and Acronyms

ICISS LOS SRR TBI

⫽ ⫽ ⫽ ⫽

ICD-9 Injury Severity Score length of stay survival risk ratio traumatic brain injury

graphic information, all discharge diagnoses, and insurance status. In contrast to the state trauma registry, all associated comorbid conditions are included, along with the ICD9-CM injury diagnoses. The entire Florida Agency for Health Care Administration data set can be used to define International Classification Injury Severity Score (ICISS) survival risk ratios (SRRs), thereby defining and controlling for the unique characteristics of the Florida population while comparing specific cohorts within that population.8,9 The SRR values for this analysis were compiled from 17 years of Florida Agency for Health Care Administration data preceding the current analysis. The 2008 and 2009 data were not used in the SRR compilation to avoid endogeneity bias. The ICISS is a robust system of evaluation that has been applied to many analyses of trauma system function and was used for risk stratification in the legislatively mandated comprehensive assessment of the Florida trauma system.10,11 This is one example of many recent reports that have defined a statistically significant survival advantage provided by state-designated trauma centers.5,12-15 The intent of this study was to determine if race, ethnicity, or insurance status played any role in access to emergency care and clinical outcome in the Florida trauma system. Motor vehicular trauma, the state’s most common and costly injury mechanism, was used to minimize other injury-related confounding variables. Logistic regression analysis of pooled data from 2008 and 2009 was used to identify factors potentially contributing to outcomes from vehicular trauma. These were then tested using chi-square analysis of the calculated model estimate, accepting p ⬍ 0.05 as significant. Specific demographic factors included age, sex, race, ethnicity, and insurance status. An ICISS survival probability ⬍ 0.85 (expected mortality rate of at least 15%) was used to define “severe” for inclusion in the study cohort. The model included 1, 2, or 3 or more comorbidities, as well as specific injury diagnoses involving traumatic brain injury (TBI), spinal cord injury, torso injury, or vascular injury. Finally, the model also included fixed effects for different types of vehicular accidents to control for inherent differences. Outcome was defined as mortality during acute hospitalization for treatment of severe injury. To control for nonsurvivable injury, the study group was further stratified by survival time after trauma center arrival to define death within 24 hours

Tepas et al

Insurance Status Determines Injury Outcomes

723

(length of stay [LOS] ⬍ 1) or thereafter (LOS ⱖ 1). To assess the likelihood that injury severity was the primary outcome determinant of the LOS ⬍ 1 cohort, the distribution of the ICISS probabilities of fatalities was analyzed. Finally the relationship between severity of injury in this group was evaluated against insurance status.

RESULTS During 2008 to 2009, 8,758 motor vehicle crash victims were treated in state-designated trauma centers. Table 1 defines the results of the entire cohort and the subcohorts whose trauma LOS was less than or beyond 24 hours. Age, sex, and injury severity were consistent predictors of survival across all timelines. Not surprisingly, mortality risk after 24 hours was driven by TBI and comorbidities. The presence of 2 or more comorbidities was a significant predictor of outcomes in this group. The absence of this effect in the LOS ⬍ 1 group probably reflects that those patients dying rapidly from their injuries did not have adequate time for diagnosis or development of comorbid conditions and injury-related complications. Race was not associated with increased mortality risk. On the contrary, within the group that survived longer than 24 hours, AfricanAmerican patients experienced reduced mortality. Being uninsured, however, was significantly related to death but only in the early mortality subgroup that died within 24 hours of admission to the trauma center (p ⬍ 0.001). Of the uninsured who died within this 24-hour period, 82% had an ICISS score (probability) of 0.5 or less, with almost 50% presenting with an ICISS score below 0.2. The relationship of injury severity and insurance status is illustrated in Figure 1. The consistently positive bars above the x-axis indicate that the uninsured had more representation in the most severe injury categories. Injury mechanism and race varied by insurance status, but the variation did not contribute significantly to the outcomes. The definitions of the vehicular injury mechanism codes are shown in Table 2. The control mechanism in the estimations was the most commonly occurring type, E812, or vehicle versus vehicle collision. Two of the injury mechanisms, car versus train and re-entrant collision, were associated with increased mortality; however, in practice they accounted for less than one-fifth of a percent of all motor vehicle events. Table 3 further defines the breakdown for vehicle–pedestrian injury, E814. It demonstrates that a lower percentage of pedestrians are uninsured. In regard to mortality, 15.5% of the 716 insured pedestrians died compared with 18.6% of the uninsured pedestrians (p ⫽ NS). Of the 35 insured pedal cyclists, 22.9% died versus 16.7% of the 12 uninsured pedal cyclists.

724

Tepas et al

Insurance Status Determines Injury Outcomes

Table 1. Logistic Regression Results Restrictions on ICISS Restrictions on LOS n Died

⬍0.85 None 8,758 941

J Am Coll Surg

⬍0.85 ⬎0 8,521 738

⬍0.85 ⬎1 8,037 529

Parameter

Estimate

Pr > ChiSq

Estimate

Pr > ChiSq

Estimate

Pr > ChiSq

Intercept Age Female African American Hispanic Other nonwhite Uninsured Medicaid E810 E811 E813 E814 E815 E816 E817 E818 E819 ICISS comorbidity1 comorbidity2 comorbidity3plus SSCI TBI Torso Vascular

2.361 0.003 ⫺0.372* ⫺0.129 ⫺0.264 0.196 0.508* ⫺0.192 1.417 3.329* ⫺0.248 0.238 0.321 0.043 ⫺0.225 0.386 0.097 ⫺7.194* ⫺0.267 ⫺0.005 0.610* ⫺0.781* ⫺0.448* ⫺0.492* ⫺0.564*

⬍0.0001 0.117 ⬍0.0001 0.365 0.051 0.253 ⬍0.0001 0.269 0.075 0.049 0.275 0.074 0.067 0.758 0.808 0.186 0.444 ⬍0.0001 0.074 0.972 0.001 ⬍0.0001 0.006 ⬍0.0001 ⬍0.0001

1.778 0.001 ⫺0.465* ⫺0.268 ⫺0.195 0.279 0.312* ⫺0.073 1.604* 3.563* ⫺0.283 0.219 0.251 0.032 0.022 0.522 0.063 ⫺6.854* ⫺0.067 0.205 0.927* ⫺0.571* ⫺0.014 ⫺0.549* ⫺0.763*

⬍0.0001 0.611 ⬍0.0001 0.095 0.178 0.126 0.030 0.683 0.048 0.043 0.256 0.128 0.197 0.830 0.981 0.084 0.649 ⬍0.0001 0.668 0.141 ⬍0.0001 0.008 0.947 ⬍0.0001 ⬍0.0001

0.611 0.003 ⫺0.601* ⫺0.365 ⫺0.224 0.189 0.291 0.128 1.853* 3.925* ⫺0.423 ⫺0.027 0.128 ⫺0.131 0.441 0.313 0.042 ⫺6.340* 0.081 0.402* 1.128* 0.036 0.628* ⫺0.562* ⫺0.556*

0.095 0.234 ⬍0.0001 0.047 0.178 0.364 0.080 0.501 0.020 0.043 0.133 0.873 0.559 0.448 0.623 0.370 0.784 ⬍0.0001 0.634 0.007 ⬍0.0001 0.904 0.029 ⬍0.0001 0.003

*Values that are statistically significant in the model (p⬍.05). E810, car vs train; E811, re-entrant collision; E813, vehicle vs vehicle; E814, vehicle vs pedestrian; E815, other vehicle vs vehicle on highway; E816, vehicle loss of control not involving other vehicle on highway; E817, Noncollision motor vehicle while boarding or alighting; E818, other vehicle accident; E819, motor vehicle of unspecified nature; ICISS, ICD-9 Injury Severity Score; LOS, length of stay; Pr, probablility; SSCI, spinal cord injury; TBI, traumatic brain injury.

DISCUSSION In a country that does not provide basic health care services for all of its citizens, and relies primarily on the entrepreneurial model of production for that provision, disparity in both access to care and therapeutic outcomes is inevitable. This has been demonstrated in numerous studies and usually addresses issues related to services needed in the convalescent or rehabilitative phases of care. Haas and Goldman16 evaluated trauma care in Massachusetts and demonstrated that uninsured patients were just as likely to be provided intensive care services as those with insurance. The uninsured, however, encountered clear barriers to operative care and physical therapy, as well as significantly higher odds of dying in hospital. Marquez de la Plata and colleagues,17 analyzing a single institution’s experience with rehabilitation placement after TBI, also dem-

onstrated that lack of insurance was a significant barrier to rehabilitation services and was associated with worse outcomes. In addition, the extent to which race contributed to worse outcomes from TBI was described in direct relation to African Americans by Hart and coworkers4 and indirectly by Staudenmayer and associates,18 citing an ethnic group consisting of multiple races that were nonwhite and non-Hispanic. The purpose of our investigation was to determine whether any of these factors were apparent in the emergency and acute phases of function of what is now a mature trauma system that has been functioning for more than 25 years. Trauma systems are designed to control the entire spectrum of the disease of injury, from prevention to societal reintegration. The emergency medical phase involves immediate response to patients with significant injury and

Vol. 212, No. 4, April 2011

Insurance Status Determines Injury Outcomes

Tepas et al

Figure 1. Difference in percent of insured and uninsured by International Classification Injury Severity Score interval. The bar above zero indicates more uninsured in the specific survival probability category.

their expeditious transport to appropriately designated trauma centers. Field assessment protocols are deliberately designed to err on the side of overtriage to minimize potential for preventable death. In trauma centers, highquality care is provided in response to determined patient need. These decisions are independent of race, ethnicity, or insurance status. This analysis of the most frequent and costly injury mechanisms encountered in Florida provides no evidence that these acute care decisions are predicated on anything beyond perceived patient need. Despite this, there is a significant difference in immediate fatality that is related to patient insurance status, suggesting that there are, in all likelihood, behavioral, socioeconomic, and physiologic determinants of outcome. The higher mortality risk is due to increased injury severity as demonstrated in Figure 1. The uninsured are more likely to be severely injured enough to die rapidly from their injuries despite the best efforts of an organized trauma system. This disparity suggests that the absence of insurance is a surrogate for presence of patient behaviors and envi-

725

ronmental factors that predispose to lethal injury, both of which may be considered characteristics of ethnicity. One potential explanation is that there is more risktaking behavior, especially as related to compliance with safety practices and licensing regulations, in uninsured patients. What is of interest is that this insurance factor disappears as an outcome determinant in the LOS ⬎ 1 cohort and is replaced by presence of 2 or more comorbidities and TBI. Given the absence of any other organ injuries in the model to emerge as significant determinants, it is most likely that the high initial mortality of the uninsured was related to lethal hypoxia, exsanguination, and/or CNS injury and that those who survived longer than 24 hours were at higher risk to develop the comorbid conditions attendant to critical care of TBI. If such is the case, then the absence of insurance reflects a risk-taking mindset that manifests itself with a higher proclivity for severe injury that is rapidly fatal or more costly to treat over an extended time. These findings differ somewhat from those of Haider and colleagues,3 who used the National Trauma Databank to analyze the effect of insurance status on adult patients stratified by race. Their findings indicated that both race and insurance status were determinants of outcome and that the uninsured sustained a higher mortality, regardless of race. They included both blunt and penetrating injury and noted a 10-fold increase in prevalence of penetrating trauma in uninsured African Americans than uninsured whites. We approached this question with a tighter focus on injury type (vehicular) and with no preset stratification of any of the logistic model’s components. Using a statewide administrative database, we avoided the issue of missing insurance data, which Haider and colleagues3 cited as a problem in 13% of their records. Injury stratification using ICISS accounts for anatomic injury and associated physiologic derangement to produce a continuous variable be-

Table 2. Distribution by Injury Mechanism, Insurance Status, and Race All

Observations, n Accident type, % Car vs train (E810) Re-entrant collision (E811) Other vehicle vs vehicle (E812) Vehicle vs vehicle (E813) Vehicle vs pedestrian (E814) Other vehicle vs vehicle on highway (E815) Vehicle loss of control not involving other vehicle on highway (E816) Noncollision motor vehicle while boarding or aligning (E817) Other vehicle accident (E818) Motor vehicle of unspecified nature (E819)

8,758 0.15 0.02 45.03 4.52 10.60 6.87 14.73 0.21 2.23 16.00

Uninsured

1,161 0.00 0.00 39.36 7.41 13.95 5.68 14.38 0.09 4.48 14.90

African American uninsured

190 0.00 0.00 36.84 7.37 18.42 8.95 11.58 0.00 4.74 12.63

726

Tepas et al

Insurance Status Determines Injury Outcomes

J Am Coll Surg

Table 3. Breakdown of Vehicle Versus Pedestrian (E814) Codes n Fourth digit designator

%

Insured

Uninsured

Insured

Uninsured

3 2 2 0 0 2 35 8 716 111 3 3

1 1 3 0 0 0 12 2 145 27 0 0

0.3 0.2 0.2 0 0 0.2 4.0 22.9 80.9 15.5 0.3 0.3

0.5 0.5 1.6 0 0 0 6.3 16.7 75.9 18.6 0 0

0 Driver of motor vehicle other than motorcycle 1 Passenger in motor vehicle other than motorcycle 2 Motorcyclist 3 Passenger on motorcycle 4 Occupant of streetcar 5 Rider of animal; occupant of animal-drawn vehicle 6 Pedal cyclist Mortality of pedal cyclists 7 Pedestrian Mortality of pedestrians 8 Other specified person 9 Unspecified person

tween 0 and 1. Moreover, by using the Florida population to determine the SRRs for each anatomic diagnosis, the comparison uses Florida-specific injury data to evaluate the factors in this regression model. Despite the differences in approach, the impact of absent insurance coverage is clearly apparent in both studies. Our findings also contrast with those of Greene and colleagues,19 who reported racial disparities in injury outcomes for patients injured in motorcycle crashes, even though African-American patients were more likely to use helmets than the white patients. Rosen and coworkers6 approached this issue of disparity in the pediatric population from the perspective that insurance coverage is a reflection of both socioeconomic status and state of health. These investigators analyzed 174,921 National Trauma Databank patients younger than 17 years, stratifying them by commercial insurance, public insurance (Medicaid), and no insurance. They also used logistic regression to demonstrate increased adjusted mortality odds for African Americans and both the uninsured and publicly insured. They postulated that insurance status was a marker for both socioeconomic and general health status. A potential limitation of this study is its application of administrative data versus that collected in a designated trauma registry. Despite this, the fact that our findings regarding insurance status and outcome are consonant with the works of Haider and colleagues,3 Rosen and coworkers,6 Greene and associates,19 and Crompton and colleagues,1 all of which were based on National Trauma Databank data, suggests that this is a very real social phenomenon. Like all of the other investigations of this issue, our study also suffers from the poorly defined distinction between race and ethnicity. Racial distinction is a biologic determinant of nature, whereas ethnicity is the result of environmental nurture. A child born in Japan and raised in the US will talk and probably behave like

his/her American peers, yet will always be Asian. Clearly, as additional research into methods to address this insurance-related phenomenon go forward, better data to define ethnicity from the perspective of nurture will be critical in changing many health-related perceptions in segments of society that are often racially diverse but share the common ethnic characteristics of poverty and limited education.20 Regardless of the significant improvements that can be anticipated to result from expansion of insurance coverage, a comprehensive solution to the problem of injury outcome disparities based on ethnicity and insurance status must include better public education, improved engineering controls, wise legislative efforts, and law enforcement. Our data indicate that the process of emergency trauma care is driven by patient pathology. The outcomes of the disease, however, are affected by lack of medical insurance. More research is needed to clarify the origins of the effects of insurance status on immediate fatality and greater injury severity. The fact that lack of insurance coverage remains a determinant of outcomes clearly indicates that improvement in management of the disease of injury depends as much on control of societal and environmental factors as it does on continued discovery of major therapeutic advances.

Author Contributions Study conception and design: Tepas, Flint, Pracht Acquisition of data: Pracht, Orban Analysis and interpretation of data: Pracht, Tepas, Flint, Orban Drafting of manuscript: Tepas Critical revision: Flint, Pracht, Orban

Vol. 212, No. 4, April 2011

REFERENCES 1. Crompton JG, Pollack KM, Oyetunji T, et al. Racial disparities in motorcycle-related mortality: an analysis of the National Trauma Data Bank. Am J Surg 2010;200:191–196. 2. Dillingham TR, Pezzin LE, Mackenzie EJ. Racial differences in the incidence of limb loss secondary to peripheral vascular disease: a population-based study. Arch Phys Med Rehabil 2002; 83:1252–1257. 3. Haider AH, Chang DC, Efron DT, et al. Race and insurance status as risk factors for trauma mortality. Arch Surg 2008;143: 945–949. 4. Hart T, O’Neil-Pirozzi TM, Williams KD, et al. Racial differences in caregiving patterns, caregiver emotional function, and sources of emotional support following traumatic brain injury. J Head Trauma Rehabil 2007;22:122–131. 5. Mackenzie EJ, Rivara FP, Jurkovich GJ, et al. The national study on costs and outcomes of trauma. J Trauma 2007;63:S54–S67; discussion S81–S86. 6. Rosen H, Saleh F, Lipsitz SR, et al. Lack of insurance negatively affects trauma mortality in US children. J Pediatr Surg 2009;44: 1952–1957. 7. Maybury RS, Bolorunduro OB, Villegas C, et al. Pedestrians struck by motor vehicles further worsen race- and insurance-based disparities in trauma outcomes: the case for inner-city pedestrian injury prevention programs. Surgery 2010;148:202–208. 8. Osler T, Rutledge R, Deis J, Bedrick E. ICISS: an international classification of disease-based injury severity score. J Trauma 1996;41:380–388. 9. Osler TM, Rogers FB, Glance LG, et al. Predicting survival, length of stay, and cost in the surgical intensive care unit: APACHE II versus ICISS. J Trauma 1998;45:234–237; discussion 237–238. 10. Celso B, Tepas J, Langland-Orban B, et al. A systematic review and meta-analysis comparing outcome of severely injured patients treated in trauma centers following the establishment of trauma systems. J Trauma 2006;60:371–378; discussion 378. 11. Durham R, Pracht E, Orban B, et al. Evaluation of a mature trauma system. Ann Surg 2006;243:775–783; discussion 783– 785. 12. MacKenzie EJ. Review of evidence regarding trauma system effectiveness resulting from panel studies. J Trauma 1999;47: S34–S41. 13. Mullins RJ, Mann NC, Hedges JR, et al. Preferential benefit of implementation of a statewide trauma system in one of two adjacent states. J Trauma 1998;44:609–616; discussion 617. 14. Pracht EE, Tepas JJ 3rd, Celso BG, et al. Survival advantage associated with treatment of injury at designated trauma centers: a bivariate probit model with instrumental variables. Med Care Res Rev 2007;64:83–97. 15. MacKenzie EJ, Rivara FP, Jurkovich GJ, et al. A national evaluation of the effect of trauma-center care on mortality. N Engl J Med 2006;354:366–378. 16. Haas JS, Goldman L. Acutely injured patients with trauma in Massachusetts: differences in care and mortality, by insurance status. Am J Public Health 1994;84:1605–1608. 17. Marquez de la Plata C, Hewlitt M, de Oliveira A, et al. Ethnic differences in rehabilitation placement and outcome after TBI. J Head Trauma Rehabil 2007;22:113–121. 18. Staudenmayer KL, Diaz-Arrastia R, de Oliveira A, et al. Ethnic disparities in long-term functional outcomes after traumatic brain injury. J Trauma 2007;63:1364–1369.

Tepas et al

Discussion

727

19. Greene WR, Oyetunji TA, Bowers U, et al. Insurance status is a potent predictor of outcomes in both blunt and penetrating trauma. Am J Surg 2010;199:554–557. 20. Smith WR, Betancourt JR, Wynia MK, et al. Recommendations for teaching about racial and ethnic disparities in health and health care. Ann Intern Med 2007;147:654–665.

Discussion DR EDWARD CORNWELL (Washington, DC): I applaud the authors for the collaboration of clinical scientists with colleagues from the Department of Health, Policy and Management in adding to the growing body of knowledge on outcomes disparities after critical injury. The challenge to get this right is substantial. The consequences of misinterpretation of data on a topic that deals with unexpected early death, race, insurance status, and access to health care are immense. Yet I am convinced that we, as academic surgeons, who, in our weekly mortality and morbidity conferences, deal with issues of preventability of death and complications in a frankly honest but collegial fashion , have just the right cultural experience to accomplish the task. Indulge me for a minute as I describe a personal event in trying to address this issue and the challenges of multivariate analysis in predicting outcomes. About 8 months ago, I sustained a basketball injury to my right eye, an eye that had a cornea transplant 35 years earlier. I was 15 minutes away from an institution well known to me, at which I had worked for 10 years. I knew about their eye center. And with a towel over my right eye, I drove myself to that center. I was met in the emergency department parking lot by a friend and former colleague whom I had recruited into the division of trauma there. He had already talked to the ophthalmology fellow. And in fact, he took my car and drove it to the parking lot, leaving me to walk in, where I was tended to by friends and former colleagues, from registration, nurses, an emergency medicine colleague, and a fellow who had the cornea specialist come in from home, who had me in the operating room a little more than 1 hour after my injury. They subsequently found that I had dehiscence of 35% of my corneal graft, lost my lens somewhere out on the basketball court, had prolapse of my iris, and later discovered that I had a detached retina. The cornea specialist and subsequent care provided to me by an anterior chamber specialist, followed by a posterior chamber specialist, over a period of 6 months and 3 operations, afforded me retention of my vision, my lifestyle, and my occupation as a surgeon. As I think of multivariate analysis— and I am intimately familiar with the Maryland statewide trauma registry— it would suggest that my race, my age greater than 50, and my sex would put me at greater risk for poorer outcomes after injury. And as I think of this story, I was provided phenomenal access because of my colleagues, because of what I knew of the system, ie, issues dealing with the provider, the patient, and the system. I was provided phenomenal access to a complicated health care system. It’s hard for me to ferret out which factors deserve the bull share of the credit for the excellent outcomes that required 3 operations and