Validating the Matching of Patients in the Linkage of a Large Hospital System's EHR with State and National Death Databases.

BACKGROUND: Though electronic health record (EHR) data have been linked to national and state death registries, such linkages have rarely been validated for an entire hospital system's EHR. OBJECTIVES: The aim of the study is to validate West Virginia University Medicine's (WVU Medicine) linkage of its EHR to three external death registries: the Social Security Death Masterfile (SSDMF), the national death index (NDI), the West Virginia Department of Health and Human Resources (DHHR). METHODS: Probabilistic matching was used to link patients to NDI and deterministic matching for the SSDMF and DHHR vital statistics records (WVDMF). In subanalysis, we used deaths recorded in Epic (n = 30,217) to further validate a subset of deaths captured by the SSDMF, NDI, and WVDMF. RESULTS: Of the deaths captured by the SSDMF, 59.8 and 68.5% were captured by NDI and WVDMF, respectively; for deaths captured by NDI this co-capture rate was 80 and 78%, respectively, for the SSDMF and WVDMF. Kappa statistics were strongest for NDI and WVDMF (61.2%) and NDI and SSDMF (60.6%) and weakest for SSDMF and WVDMF (27.9%). Of deaths recorded in Epic, 84.3, 85.5, and 84.4% were captured by SSDMF, NDI, and WVDMF, respectively. Less than 2% of patients' deaths recorded in Epic were not found in any of the death registries. Finally, approximately 0.2% of "decedents" in any death registry re-emerged in Epic at least 6 months after their death date, a very small percentage and thus further validating the linkages. CONCLUSION: NDI had greatest validity in capturing deaths in our EHR. As a similar, though slightly less capture and agreement rate in identifying deaths is observed for SSDMF and state vital statistics records, these registries may be reasonable alternatives to NDI for research and quality assurance studies utilizing entire EHRs from large hospital systems. Investigators should also be aware that there will be a very tiny fraction of "dead" patients re-emerging in the EHR.

Mortality and Associated Morbidities Following Traumatic Brain Injury in Older Medicare Statin Users.

OBJECTIVE: To assess the relationship between posttraumatic brain injury statin use and (1) mortality and (2) the incidence of associated morbidities, including stroke, depression, and Alzheimer's disease and related dementias following injury. SETTING AND PARTICIPANTS: Nested cohort of all Medicare beneficiaries 65 years of age and older who survived a traumatic brain injury (TBI) hospitalization during 2006 through 2010. The final sample comprised 100 515 beneficiaries. DESIGN: Retrospective cohort study of older Medicare beneficiaries. Relative risks (RR) and 95% confidence interval (CI) were obtained using discrete time analysis and generalized estimating equations. MEASURES: The exposure of interest included monthly atorvastatin, fluvastatin, lovastatin, pravastatin, rosuvastatin, and simvastatin use. Outcomes of interest included mortality, stroke, depression, and Alzheimer's disease and related dementias. RESULTS: Statin use of any kind was associated with decreased mortality following TBI hospitalization discharge. Any statin use was also associated with a decrease in any stroke (RR, 0.86; 95% confidence intervals (CI), 0.81-0.91), depression (RR, 0.85; 95% CI, 0.79-0.90), and Alzheimer's disease and related dementias (RR, 0.77; 95% CI, 0.73-0.81). CONCLUSION: These findings provide valuable information for clinicians treating older adults with TBI as clinicians can consider, when appropriate, atorvastatin and simvastatin to older adults with TBI in order to decrease mortality and associated morbidities.

Anxiety and Posttraumatic Stress Disorder Among Medicare Beneficiaries After Traumatic Brain Injury.

OBJECTIVE: To estimate rates of anxiety and posttraumatic stress disorder (PTSD) diagnoses after traumatic brain injury (TBI) among Medicare beneficiaries, quantify the increase in rates relative to the pre-TBI period, and identify risk factors for diagnosis of anxiety and PTSD. PARTICIPANTS: A total of 96 881 Medicare beneficiaries hospitalized with TBI between June 1, 2006 and May 31, 2010. DESIGN: Retrospective cohort study. MEASURES: Diagnosis of anxiety (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 300.0x) and/or PTSD (ICD-9-CM code 309.81). RESULTS: After TBI, 16 519 (17%) beneficiaries were diagnosed with anxiety and 269 (0.3%) were diagnosed with PTSD. Rates of anxiety and PTSD diagnoses were highest in the first 5 months post-TBI and decreased over time. Pre-TBI diagnosis of anxiety disorder was significantly associated with post-TBI anxiety (risk ratio, 3.55; 95% confidence interval, 3.42-3.68) and pre-TBI diagnosis of PTSD was significantly associated with post-TBI PTSD (risk ratio 70.09; 95% confidence interval 56.29-111.12). CONCLUSION: This study highlights the increased risk of anxiety and PTSD after TBI. Routine screening for anxiety and PTSD, especially during the first 5 months after TBI, may help clinicians identify these important and treatable conditions, especially among patients with a history of psychiatric illness.

Effect of Maryland's 2011 Alcohol Sales Tax Increase on Alcohol-Positive Driving.

INTRODUCTION: The 2011 Maryland alcohol sales tax increase from 6% to 9% provided an opportunity to evaluate the impact on rates of alcohol-positive drivers involved in injury crashes. METHODS: Maryland police crash reports from 2001 to 2013 were analyzed using an interrupted time series design and a multivariable analysis employing generalized estimating equations models with a negative binomial distribution. Data were analyzed in 2014-2015. RESULTS: There was a significant gradual annual reduction of 6% in the population-based rate of all alcohol-positive drivers (p<0.03), and a 12% reduction for drivers aged 15-20 years (p<0.007), and 21-34 years (p<0.001) following the alcohol sales tax increase. There were no significant changes in rates of alcohol-positive drivers aged 35-54 years (rate ratio, 0.98; 95% CI=0.89, 1.09). Drivers aged ≥55 years had a significant immediate 10% increase in the rate of alcohol-positive drivers (rate ratio, 1.10; 95% CI=1.04, 1.16) and a gradual increase of 4.8% per year after the intervention. Models using different denominators and controlling for multiple factors including a proxy for unmeasured factors found similar results overall. CONCLUSIONS: The 2011 Maryland alcohol sales tax increase led to a significant reduction in the rate of all alcohol-positive drivers involved in injury crashes especially among drivers aged 15-34 years. This is the first study to examine the impact of alcohol sales taxes on crashes; previous research focused on excise tax. Increasing alcohol taxes is an important but often neglected intervention to reduce alcohol-impaired driving.

Treatment Charges for Traumatic Brain Injury Among Older Adults at a Trauma Center.

OBJECTIVE: To provide charge estimates of treatment for traumatic brain injury (TBI), including both hospital and physician charges, among adults 65 years and older treated at a trauma center. METHODS: We identified older adults treated for TBI during 2008-2012 (n = 1843) at Maryland's Primary Adult Resource Center and obtained hospital and physician charges separately. Analyses were stratified by sex and all charges were inflated to 2012 dollars. Total TBI charges were modeled as a function of covariates using a generalized linear model. RESULTS: Women comprised 48% of the sample. The mean unadjusted total TBI hospitalization charge for adults 65 years and older was $36 075 (standard deviation, $63 073). Physician charges comprised 15% of total charges. Adjusted mean charges were lower in women than in men (adjusted difference, -$894; 95% confidence interval, -$277 to -$1512). Length of hospital and intensive care unit stay were associated with the highest charges. CONCLUSIONS: This study provides the first estimates of hospital and physician charges associated with hospitalization for TBI among older adults at a trauma center that will aid in resource allocation, triage decisions, and healthcare policy.

Psychotropic Medication Use among Medicare Beneficiaries Following Traumatic Brain Injury.

OBJECTIVES: To characterize psychotropic medication use before and after traumatic brain injury (TBI) hospitalization among older adults. A secondary objective is to determine how receipt of indicated pharmacologic treatment for anxiety and post-traumatic stress disorder (PTSD) differs following TBI. DESIGN: Retrospective cohort. SETTING: United States. PARTICIPANTS: Medicare beneficiaries aged ≥65 years hospitalized with TBI between 2006 and 2010 with continuous drug coverage for 12 months before and after TBI (N = 60,276). MEASUREMENTS: We obtained monthly psychotropic medication use by drug class and specific drugs from Medicare Part D drug event files.ICD-9 codes were used to define anxiety (300.0x) and PTSD (309.81). RESULTS: Average monthly prevalence of psychotropic medication use among all patients hospitalized for TBI was 44.8%; antidepressants constituted 73%. Prevalence of psychotropic medication use increased from 2006 to 2010. Following TBI, psychotropic medication use increased slightly (OR: 1.05; 95% CI: 1.03, 1.06.) Tricyclic antidepressant use decreased post-TBI (OR: 0.76; 95% CI: 0.73, 0.79) whereas use of the sedating antidepressants mirtazapine (OR: 1.31; 95% CI: 1.25, 1.37) and trazadone (OR: 1.11; 95% CI: 1.06, 1.17) increased. Antipsychotic (OR: 1.15; 95% CI: 1.12, 1.19) use also increased post-TBI. Beneficiaries newly diagnosed with anxiety (OR: 0.42; 95% CI: 0.36, 0.48) and/or PTSD (OR: 0.39; 95% CI: 0.18, 0.84) post-TBI were less likely to receive indicated pharmacologic treatment. CONCLUSIONS: Older adults hospitalized with TBI have a high prevalence of psychotropic medication use yet are less likely to receive indicated pharmacological treatment for newly diagnosed anxiety and PTSD following TBI.

Can progress in reducing alcohol-impaired driving fatalities be resumed? Results of a workshop sponsored by the Transportation Research Board, Alcohol, Other Drugs, and Transportation Committee (ANB50).

OBJECTIVE: Despite successes in the 1980s and early 1990s, progress in reducing impaired driving fatalities in the United States has stagnated in recent years. Since 1997, the percentage of drivers involved in fatal crashes with illegal blood alcohol concentration (BAC) levels has remained at approximately 20 to 22%. Many experts believe that public complacency, competing social and public health issues, and the lack of political fortitude have all contributed to this stagnation. The number of alcohol-related crashes, injuries, and fatalities is still unacceptable, and most are preventable. The public needs to be aware that the problem presented by drinking drivers has not been solved. Political leaders need guidance on which measures will affect the problem, and stakeholders need to be motivated once again to implement effective strategies. METHODS: The National Academy of Sciences (NAS) Transportation Research Board (TRB), Alcohol, Other Drugs, and Transportation Committee (ANB50) sponsored a workshop held at the NAS facility in Woods Hole, Massachusetts, on August 24-25, 2015, to discuss the lack of progress in reducing impaired driving and to make recommendations for future progress. A total of 26 experts in research and policy related to alcohol-impaired driving participated in the workshop. The workshop began by examining the static situation in the rate of alcohol-impaired driving fatal crashes to determine what factors may be inhibiting further progress. The workshop then discussed 8 effective strategies that have not been fully implemented in the United States. Workshop participants (16 of the 26) rated their top 3 strategies. RESULTS: 3 strategies received the most support: 1. Impose administrative sanctions for drivers with BACs = 0.05 to 0.08 g/dL. 2. Require alcohol ignition interlocks for all alcohol-impaired driving offenders. 3. Increase the frequency of sobriety checkpoints, including enacting legislation to allow them in the 11 states that currently prohibit them. 5 other important strategies included the following: (1) increase alcohol taxes to raise the price and reduce alcohol consumption; (2) reengage the public and raise the priority of impaired driving; (3) lower the illegal per se BAC limit to 0.05 for a criminal offense; (4) develop and implement in-vehicle alcohol detection systems; and (5) expand the use of screening and brief interventions in medical facilities. CONCLUSIONS: Each of these strategies is proven to be effective, yet all are substantially underutilized. Each is used in some jurisdictions in the United States or Canada, but none is used extensively. Any one of the 3 strategies implemented on a widespread basis would decrease impaired driving crashes, injuries, and fatalities. Based on the research, all 3 together would have a substantial impact on the problem.

Patterns of Depression Treatment in Medicare Beneficiaries with Depression after Traumatic Brain Injury.

There are no clinical guidelines addressing the management of depression after traumatic brain injury (TBI). The objectives of this study were to (1) describe depression treatment patterns among Medicare beneficiaries with a diagnosis of depression post-TBI; (2) compare them with depression treatment patterns among beneficiaries with a diagnosis of depression pre-TBI; and (3) quantify the difference in prevalence of use. We conducted a retrospective analysis of Medicare beneficiaries hospitalized with TBI during 2006-2010. We created two cohorts: beneficiaries with a new diagnosis of depression pre-TBI (n=4841) and beneficiaries with a new diagnosis of depression post-TBI (n=4668). We searched for antidepressant medications in Medicare Part D drug event files and created variables indicating antidepressant use in each 30-day period after diagnosis of depression. We used provider specialty and current procedural terminology to identify psychotherapy in any location. We used generalized estimating equations to quantify the effect of TBI on receipt of depression treatment during the year after diagnosis of depression. Average monthly prevalence of antidepressant use was 42% among beneficiaries with a diagnosis of depression pre-TBI and 36% among those with a diagnosis post-TBI (p<0.001). Beneficiaries with a diagnosis of depression post-TBI were less likely to receive antidepressants compared with a depression diagnosis pre-TBI (adjusted odds ratio [OR] 0.87; 95% confidence interval [CI] 0.82, 0.92). There was no difference in receipt of psychotherapy between the two groups (OR 1.08; 95% CI 0.93, 1.26). Depression after TBI is undertreated among older adults. Knowledge about reasons for this disparity and its long-term effects on post-TBI outcomes is limited and should be examined in future work.

Are current law enforcement strategies associated with a lower risk of repeat speeding citations and crash involvement? A longitudinal study of speeding Maryland drivers.

PURPOSE: To determine whether traffic court appearances and different court verdicts were associated with risk of subsequent speeding citations and crashes. METHODS: A cohort of 29,754 Maryland drivers ticketed for speeding who either went to court or paid fines by mail in May/June 2003 was followed for 3 years. Drivers appearing in court were categorized by verdicts: 1) not guilty, 2) suspension of prosecution/no prosecution (STET/NP), 3) case dismissed, 4) probation before judgment (PBJ) and fines, or 5) fines and demerit points. Cox proportional hazard models were used to estimate adjusted hazard ratios (AHR). RESULTS: Court appearances were associated with lower risk of subsequent speeding citations (AHR = 0.92; 95% confidence interval [CI], 0.88-0.96), but higher risk of crashes (AHR = 1.25; 95% CI, 1.16-1.35). PBJ was associated with significantly lower repeat speeding tickets (AHR = 0.83; 95% CI, 0.75-0.91) and a non-significant decrease in crashes (AHR = 0.87; 95% CI, 0.75-1.02). Both repeat speeding tickets and subsequent crashes were significantly lower in the STET/NP group. CONCLUSIONS: PBJ and STET/NP may reduce speeding and crashes, but neither verdict eliminated excess crash risk among drivers who choose court appearances. Randomized, controlled evaluations of speeding countermeasures are needed to inform traffic safety policies.

Etiology of work-related electrical injuries: a narrative analysis of workers' compensation claims.

The purpose of this study was to provide new insight into the etiology of primarily nonfatal, work-related electrical injuries. We developed a multistage, case-selection algorithm to identify electrical-related injuries from workers' compensation claims and a customized coding taxonomy to identify pre-injury circumstances. Workers' compensation claims routinely collected over a 1-year period from a large U.S. insurance provider were used to identify electrical-related injuries using an algorithm that evaluated: coded injury cause information, nature of injury, "accident" description, and injury description narratives. Concurrently, a customized coding taxonomy for these narratives was developed to abstract the activity, source, initiating process, mechanism, vector, and voltage. Among the 586,567 reported claims during 2002, electrical-related injuries accounted for 1283 (0.22%) of nonfatal claims and 15 fatalities (1.2% of electrical). Most (72.3%) were male, average age of 36, working in services (33.4%), manufacturing (24.7%), retail trade (17.3%), and construction (7.2%). Body part(s) injured most often were the hands, fingers, or wrist (34.9%); multiple body parts/systems (25.0%); lower/upper arm; elbow; shoulder, and upper extremities (19.2%). The leading activities were conducting manual tasks (55.1%); working with machinery, appliances, or equipment; working with electrical wire; and operating powered or nonpowered hand tools. Primary injury sources were appliances and office equipment (24.4%); wires, cables/cords (18.0%); machines and other equipment (11.8%); fixtures, bulbs, and switches (10.4%); and lightning (4.3%). No vector was identified in 85% of cases. and the work process was initiated by others in less than 1% of cases. Injury narratives provide valuable information to overcome some of the limitations of precoded data, more specially for identifying additional injury cases and in supplementing traditional epidemiologic data for further understanding the etiology of work-related electrical injuries that may lead to further prevention opportunities.

Work-related ladder fall fractures: identification and diagnosis validation using narrative text.

OBJECTIVE: To identify ladder-related fracture injuries and determine how ladder fall fractures differ from other ladder-related injuries. METHODS: Ladder-related fracture cases were identified using narrative text and coded data from workers' compensation claims. Potential cases were identified by text searches and verified with claim records. Injury characteristics were compared using proportionate injury ratios. RESULTS: Of 9826 ladder-related injuries, 7% resulted in fracture cases. Falls caused 89% of fractures and resulted in more medical costs and disability days than other injuries. Frequent mechanisms were ladder instability (22%) and lost footing (22%). Narrative text searches identified 17% more fractures than injury codes alone. Males were more likely to sustain a fall fracture than other injuries; construction workers were most likely, and retail workers were the least likely to sustain fractures. CONCLUSIONS: Fractures are an important injury from ladder falls, resulting more serious consequences than other ladder-related injuries. Text analysis can improve the quality and utility of workers compensation data by identifying and understanding injury causes. Proportionate injury ratios are also useful for making cross-group comparisons of injury experience when denominator data are not available. Greater attention to risk factors for ladder falls is needed for targeting interventions.

The relationship between safety climate and injury rates across industries: the need to adjust for injury hazards.

Previous studies have suggested that strong safety climates (shared perceptions of safe conducts at work) are associated with lower workplace-injury rates, but they rarely control for differences in industry hazards. Based on 33 companies, we assessed its association with injury rates using three rate based injury measures (claims per 100 employees, claims per 100,000 h worked, and claims per 1 million US dollars payroll), which were derived from workers' compensation injury claims. Linear regression models were used to test the predictability of safety climate on injury rates, followed by controlling for differences in hazard across industries gauged by national industry-specific injury rates. In the unadjusted model, company level safety climate were negatively and significantly associated with injury rates. However, all of the above associations were no longer apparent when controlling for the hazardousness of the specific industry. These findings may be due to over adjustment of hazard risk, or the overwhelming effects of industry specific hazards relative to safety climate effects that could not be differentiated with the statistical power in our study. Industry differences in hazard, conceptualized as one type of injury risk, however need to be considered when testing the association between safety climate and injury across different industries.

The use of sentinel injury deaths to evaluate the quality of multiple source reporting for occupational injuries.

PURPOSE: This study sought to develop an efficient method for evaluating the validity and completeness of routinely available sources of occupational injury fatality data. METHODS: Deaths due to falls from elevations, machinery, and electrocutions were selected as sentinel injuries likely to have occurred at work. Deaths from these injuries were identified from Maryland vital statistics over 7 years. The work-relatedness of these injuries and sensitivity of reporting were determined from death certificates, medical examiner reports, the National Traumatic Occupational Fatality System (NTOF), the Maryland Occupational Safety and Health Administration (MOSH), and Workers' Compensation (WC) data. RESULTS: A total of 527 deaths were identified for ages 16 and above, of which, 45% were work-related. Identification of work-related deaths varied by source: medical examiner (100%), death certificates (89%), NTOF (68%), MOSH (59%), and WC (44%). Reporting differed by age, cause of injury, year, occupation, and industry. CONCLUSIONS: Examination of work-relatedness for deaths from certain causes is an efficient means of evaluating the quality of occupational injury reporting source data. These sentinel injuries uncovered significant underreporting in sources used by national surveillance systems, resulted in improved NTOF reporting, and suggest the need to make more use of medical examiner data when available.

Unmet substance abuse treatment need, health services utilization, and cost: a population-based emergency department study.

STUDY OBJECTIVE: There is a high prevalence of unmet substance abuse treatment need among adult hospital emergency department (ED) patients. We examine the association between this unmet need and excess utilization of health services and estimate costs. METHODS: A statewide, 2-stage, probability sample survey was conducted in 7 Tennessee general hospital EDs from June 1996 to January 1997. Toxicologic screening augmented in-person interviews. Main outcome measures were ED case disposition; frequency of physician office visits, ED visits, and hospitalizations in the past 12 months; and costs of excess service utilization. Covariates in the multivariate model were substance abuse treatment need status, age, sex, main reason for ED visit, perceived previous health status, history of tobacco use, and health care coverage. Unmet substance abuse treatment need was assessed using 13 overlapping criteria that incorporated use, dependence, denial, and treatment history. Target substances included ethanol and selected illegal and prescription drugs but not nicotine. RESULTS: Compared with patients without substance abuse treatment need (n=1,073), patients with unmet need (n=415) were 81% more likely to be admitted to the hospital during their current ED visit (odds ratio [OR] 1.81; 95% confidence interval [CI] 1.27 to 2.64) and 46% more likely to have reported making at least 1 ED visit in the previous 12 months (OR 1.46; 95% CI 1.12 to 1.84). Their utilization patterns accounted for an estimated 777.2 million US dollars in extra hospital charges for Tennessee in 2000 dollars, representing an additional 1,568 US dollars per ED patient with unmet substance abuse treatment need. CONCLUSION: ED patients with unmet substance abuse treatment need generated much higher hospital and ED charges than patients without such need. Given potential savings from avoidable health care costs, the future burden of substance-associated ED visits and hospitalizations may be reduced through programs that screen and, as appropriate, provide brief interventions or treatment options to these patients.

Assessing substance abuse treatment need: a statewide hospital emergency department study.

STUDY OBJECTIVE: Health care providers in hospital emergency departments rarely take substance abuse histories or assess associated treatment need. This study compares documentation of psychoactive drug-related diagnoses for adult ED patients in medical records with treatment need assessed through self-report, toxicologic screening, and Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV), criteria. METHODS: A statewide, 2-stage, probability sample survey was conducted in 7 Tennessee general hospital EDs from June 1996 to January 1997. Main outcome measures were the prevalence of diagnosed substance abuse problems, positive bioassay results, denied use, and treatment need. Sensitivity and multivariate analyses were conducted by using varied case definitions of treatment need. RESULTS: Thirty-one percent (95% confidence interval [CI] 27.3% to 34.7%) of screened ED patients (n=1,330) had positive test results for substance use. Their prevalence of denial of use in the 30 days before the survey ranged from 10% for alcohol (95% CI 5.7% to 14.3%) to 100% for phencyclidine. One percent of all ED patients (n=1,502) had a recorded diagnosis of substance abuse. By contrast, as many as 27% (95% CI 23.3% to 31.8%) were assessed as needing substance abuse treatment on the basis of a comprehensive case definition that accounted for denial and positive test results. A sensitivity analysis using other case definitions is also presented. For example, 4% (95% CI 2.8% to 5.3%) of patients met the very strict definition of DSM-IV current drug dependence only. Under the comprehensive case definition, TennCare patients (adjusted odds ratio [OR] 1.63; 95% CI 1.30 to 2.05) and Medicare patients (adjusted OR 2.50; 95% CI 1.34 to 4.65) showed excess treatment need relative to the privately insured. Excess need was also exhibited by patients reporting 1 or more prior ED visits in the past year (adjusted OR 1.62; 95% CI 1.13 to 2.31) and by patients taking 2 or more hours to reach the ED after the onset of injury or illness (adjusted OR 1.54; 95% CI 1.16 to 2.04). Treatment need was inversely associated with age. Irrespective of case definition, less than 10% of ED patients who needed substance abuse treatment were receiving such treatment. CONCLUSION: EDs can be important venues for detecting persons in need of substance abuse treatment.

Socioeconomic inequalities in injury: critical issues in design and analysis.

Injuries continue to place a tremendous burden on the public's health and rates vary widely among different groups in the population. Increasing attention has recently been given to the effects of socioeconomic status (SES) as a determinant of health among both individuals and communities. However, relatively few studies have focused on the influence of SES and injuries. Furthermore, those that have, and the other injury studies that have included measures of SES in their analysis, have varying degrees of conceptual and methodological rigor in their use of this measure. Recent advances in data linkage and analytic techniques have, however, provided new and improved methods to assess the relationship between SES and injuries. This review summarizes the relevant literature on SES and injuries, with particular attention to study design, and the measurement and interpretation of SES. We found that increasing SES has a strong inverse association with the risk of both homicide and fatal unintentional injuries, although the results for suicide were mixed. However, the relationship between SES and nonfatal injuries was less consistent than for fatal injuries. We offer potential explanatory mechanisms for the relationship between SES and injuries and make recommendations for future research in this area.