Brianna Mills

Stepping towards causation in studies of neighborhood and environmental effects: How twin research can overcome problems of selection and reverse causation

No causal evidence is available to translate associations between neighborhood characteristics and health outcomes into beneficial changes to built environments. Observed associations may be causal or result from uncontrolled confounds related to family upbringing. Twin designs can help neighborhood effects studies overcome selection and reverse causation problems in specifying causal mechanisms. Beyond quantifying genetic effects (i.e., heritability coefficients), we provide examples of innovative measures and analytic methods that use twins as quasi-experimental controls for confounding by environmental effects. We conclude that collaboration among investigators from multiple fields can move the field forward by designing studies that step toward causation.

State Firearm Legislation and Nonfatal Firearm Injuries.

OBJECTIVES We investigated whether stricter state-level firearm legislation was associated with lower hospital discharge rates for nonfatal firearm injuries. METHODS We estimated discharge rates for hospitalized and emergency department-treated nonfatal firearm injuries in 18 states in 2010 and used negative binomial regression to determine whether strength of state firearm legislation was independently associated with total nonfatal firearm injury discharge rates. RESULTS We identified 26 744 discharges for nonfatal firearm injuries. The overall age-adjusted discharge rate was 19.0 per 100 000 person-years (state range = 3.3-36.6), including 7.9 and 11.1 discharges per 100 000 for hospitalized and emergency department-treated injuries, respectively. In models adjusting for differences in state sociodemographic characteristics and economic conditions, states in the strictest tertile of legislative strength had lower discharge rates for total (incidence rate ratio [IRR] = 0.60; 95% confidence interval [CI] = 0.44, 0.82), assault-related (IRR = 0.58; 95% CI = 0.34, 0.99), self-inflicted (IRR = 0.18; 95% CI = 0.14, 0.24), and unintentional (IRR = 0.53; 95% CI = 0.34, 0.84) nonfatal firearm injuries. CONCLUSIONS There is significant variation in state-level hospital discharge rates for nonfatal firearm injuries, and stricter state firearm legislation is associated with lower discharge rates for such injuries.

Facility Characteristics and Inhospital Pediatric Mortality after Severe Traumatic Brain Injury

More than 500,000 children sustain a traumatic brain injury (TBI) each year. Previous studies have described significant variation in inhospital mortality after pediatric TBI. The aim of this study was to identify facility-level characteristics independently associated with 30-day inhospital mortality after pediatric severe TBI. We hypothesized that, even after accounting for patient-level characteristics associated with mortality, the characteristics of facilities where patients received care would be associated with inhospital mortality. Using data from the National Trauma Data Bank from 2009-2012, we identified a cohort of 6707 pediatric patients hospitalized with severe TBI in 391 facilities and investigated their risk of 30-day inhospital mortality. Pre-specified facility-level characteristics (trauma certification level, teaching status, census region, facility size, nonprofit status, and responsibility for pediatric trauma care) were added to a Poisson regression model that accounted for patient-level characteristics associated with mortality. In multivariable analyses, patients treated in facilities located in the Midwest (risk ratio [RR]=1.42; 95% confidence interval [CI] 1.12-1.81) and South (RR=1.39; 95% CI: 1.12-1.72) regions had higher likelihoods of 30-day inhospital mortality compared with patients treated in the Northeast. Other facility-level characteristics were not found to be significant. To our knowledge, this is one of the largest investigations to identify regional variation in inhospital mortality after pediatric severe TBI in a national sample after accounting for individual and other facility-level characteristics. Further investigations to help explain this variation are needed to inform evidence-based decision-making for pediatric severe TBI care across different settings.

The Application of the CRASH-CT Prognostic Model for Older Adults With Traumatic Brain Injury: A Population-Based Observational Cohort Study.

Objective:To examine the performance of the Corticosteroid Randomization After Significant Head injury (CRASH) trial prognostic model in older patients with traumatic brain injury. Setting:The National Study on Costs and Outcomes of Trauma cohort, established at 69 hospitals in the United States in 2001 and 2002. Participants:Adults with traumatic brain injury and an initial Glasgow Coma Scale score of 14 or less. Design:The CRASH-CT model predicting death within 14 days was deployed in all patients. Model performance in older patients (aged 65-84 years) was compared with that in younger patients (aged 18-64 years). Main Measures:Model discrimination (as defined by the c-statistic) and calibration (as defined by the Hosmer-Lemeshow P value). Results:CRASH-CT model discrimination was not significantly different between the older (n = 356; weighted n = 524) and younger patients (n = 981; weighted n = 2602) and was generally adequate (c-statistic 0.83 vs 0.87, respectively; P = .11). CRASH-CT model calibration was adequate for the older patients and inadequate for younger patients (Hosmer-Lemeshow P values .12 and .001, respectively), possibly reflecting differences in sample size. Calibration-in-the-large showed no systematic under- or overprediction in either stratum. Conclusion:The CRASH-CT model may be valid for use in a geriatric population.

Prior Arrest, Substance Use, Mental Disorder, and Intent-Specific Firearm Injury

INTRODUCTION Substance use, mental disorders, and arrest are markers of increased firearm injury risk. It is unclear how these markers vary by intent. Examining these interrelated factors together can clarify their associations with assault-related, self-inflicted, unintentional, and legal intervention firearm injuries, informing intent-specific interventions. METHODS In 2017-2018, 2-year diagnosis and arrest histories of intent-specific firearm injury cases were compared with those of unintentionally injured motor vehicle collision passenger controls. Fatal and nonfatal firearm and motor vehicle collision injury records in Seattle (2010-2014) were linked to statewide hospitalization and arrest records. Multinomial logistic regression models compared odds of prior arrest, substance use, and mental disorder diagnoses among intent-specific firearm injury cases relative to controls, adjusting for age, race, and gender. RESULTS A total of 763 cases and 335 controls were identified. Unintentional and self-inflicted cases did not differ significantly from controls in arrest history. Legal intervention cases resembled assault-related cases in their arrest history, and self-inflicted cases in their hospitalization history. The legal intervention cases were more likely than controls to have a prior felony arrest (OR=7.72, 95% CI=2.63, 20.97), and diagnoses involving alcohol (OR=4.06, 95% CI=1.04, 15.84); cannabis (OR=11.00, 95% CI=1.01, 119.36); depression/anxiety (OR=7.22, 95% CI=1.89, 27.67); psychosis (OR=6.99, 95% CI=1.35, 36.24); or conduct disorder (OR=22.01, 95% CI=1.44, 335.93). CONCLUSIONS Individuals with intent-specific firearm injuries have distinct patterns of prior substance use, mental disorder, and arrest. Many injuries occur after a series of encounters with institutions meant to help individuals during crises that can fail to provide longer-term solutions.

Donor support for quality assurance and pharmacovigilance of anti-malarials in malaria-endemic countries

BackgroundMalaria control efforts have been strengthened by funding from donor groups and government agencies. The Global Fund to Fight AIDS, Tuberculosis and the Malaria (Global Fund), the US President’s Malaria Initiative (PMI) account for the majority of donor support for malaria control and prevention efforts. Pharmacovigilance (PV), which encompasses all activities relating to the detection, assessment, understanding, and prevention of adverse effects or any other drug-related problem, is a necessary part of efforts to reduce drug resistance and improve treatment outcomes. This paper reports on an analysis of PV plans in the Global Fund and PMI and World Bank’s grants for malaria prevention and control.MethodsAll active malaria grants as of September 2015 funded by the Global Fund and World Bank, and fiscal year 2015 and 2016 PMI Malaria Operational Plans (MOP) were identified. The total amount awarded for PV-related activities and drug quality assurance was abstracted. A Key-Word-in-Context (KWIC) analysis was conducted for the content of each grant. Specific search terms consisted of pharmacovigilance, pregn*, registry, safety, adverse drug, mass drug administration, primaquine, counterfeit, sub-standard, and falsified. Grants that mentioned PV activities identified in the KWIC search, listed PV in their budgets, or included the keywords: counterfeit, sub-standard, falsified, mass drug administration, or adverse event were thematically coded using Dedoose software version 7.0.ResultsThe search identified 159 active malaria grants including 107 Global Fund grants, 39 fiscal year 2015 and 2016 PMI grants and 13 World Bank grants. These grants were primarily awarded to low-income countries (57.2%) and in sub-Saharan Africa (SSA) (70.4%). Thirty-seven (23.3%) grants included a budget line for PV- or drug quality assurance–related activities, including 21 PMI grants and 16 Global Fund grants. Only 23 (14.5%) grants directly mentioned PV. The primary focus area was improving drug quality monitoring, especially among the PMI grants.ConclusionsThe results of the analysis demonstrate that funding for PV has not been sufficiently prioritized by either the key malaria donor organizations or by the recipient countries, as reflected in their grant proposal submissions and MOPs.

Perioperative hypotension and discharge outcomes in non-critically injured trauma patients, a single centre retrospective cohort study

BACKGROUND There is a lack of information on the effect of age on perioperative care and outcomes after minor trauma in the elderly. We examined the association between perioperative hypotension and discharge outcome among non-critically injured adult patients. METHODS We conducted a retrospective study of non-critically ill patients (ISS <9 or discharged within less than 24h) who received anaesthesia care for surgery and Recovery Room care at a level-1 trauma centre between 5/1/2012 and 11/30/2013. Perioperative hypotension was defined as systolic blood pressure (SBP) <90mmHg (traditional measure) for all patients, and SBP <110mmHg (strict measure) for patients ≥65years. Poor outcome was defined as death or discharge to skilled nursing facility/hospice. RESULTS 1744 patients with mean ISS 4.4 across age groups were included; 169 (10%) were ≥65years. Among patients≥65years, intraoperative hypotension occurred in >75% (131/169, traditional measure) and in >95% (162/169, strict measure); recovery room hypotension occurred in 2% (4/169) and 29% (49/169), respectively. Mean age-adjusted anaesthetic agent concentration (MAC) was similar across age groups. Opioid use decreased from 9.3 (SD 5.7) mg/h morphine equivalents in patients <55years to 6.2 (SD 4.0) mg/h in patients over 85 years. Adjusted for gender, ASA score, anaesthesia duration, morphine equivalent/hr, fluid balance, MAC and surgery type, and using traditional definition, older patients were more likely than patients <55 to experience perioperative hypotension: aRR 1.21, 95% CI 1.11-1.30 for 55-64 and aRR 1.19, 95% CI 1.07-1.32 for ages 65-74. Perioperative hypotension was associated with poor discharge outcome (aRR 1.55; 95% CI 1.04-2.31 and aRR 1.87; 95% CI 1.17-2.98, respectively). CONCLUSION Despite age related reduction in doses of volatile anaesthetic and opioids administered during anaesthesia care, and regardless of hypotension definition used, non-critically injured patients undergoing surgery experience a large perioperative hypotension burden. This burden is higher for patients 55-74 years and older and is a risk factor for poor discharge outcomes, independent of age and ASA status.

41 Patterns of prior hospitalisation and arrest: a comparison of legal intervention firearm injuries to assault-related firearm injuries and motor vehicle injuries in seattle

Statement of purpose Legal intervention (LI) firearm injuries may occur after contact with both the legal and medical systems; examining arrest and medical records jointly may provide clearer information on risk markers for such injuries. We examined whether LI shooting victims differed from other firearm and non-firearm injury victims in their prior contacts with the legal and medical systems. Methods We probabilistically linked 2010–2014 fatal and non-fatal LI and assault-related firearm injuries and unintentional motor vehicle crash (MVC) passenger injuries that either occurred in Seattle or were sustained by Seattle residents to statewide hospitalisation and arrest records. Multinomial logistic regression was used to compare the odds of arrest and hospitalisation for substance abuse or mental illness in the 2 years prior to injury among LI and assault-related firearm injury cases relative to MVC passenger injury controls, adjusting for age, race, and gender. Results We identified 31 LI and 442 assault-related firearm injuries, and 331 MVC passenger injuries. LI and assault-related cases were more likely than MVC controls to have a felony record (OR=6.34; 95% CI: 2.29–17.53 and OR=4.26; 95% CI: 2.39–7.59, respectively). LI cases were more likely than MVC controls to have a conduct-related (OR=22.01; 95% CI: 1.44–335.93) or psychosis diagnosis (OR=5.37; 95% CI: 1.12–25.83). LI cases were more likely than assault-related cases to have a substance abuse (OR=3.91; 95% CI: 1.14–13.47) or mental illness hospitalisation (OR=3.44; 95% CI: 1.02–11.54). Conclusion The burden of past arrest and substance abuse and mental illness hospitalisation was greater among LI firearm injury victims than among individuals with firearm assault or MVC passenger injuries. Significance Joint review of arrest and medical records can differentiate LI firearm injury victims from victims with other injuries. Legal intervention firearm injury prevention strategies can be tailored using these patterns.

20 Firearm-related injuries in-hospital settings and in-hospital mortality

Statement of purpose The proportion of injuries involving firearms varies from <1% of unintentional nonfatal injuries to 51% of suicides. Differences between firearm injuries treated in emergency department (ED) and inpatient settings are not well characterised. To ascertain how mortality varies by treatment setting and injury intent, we examined associations between patient characteristics, intent, and risk of death in ED and inpatient settings in a large sample of patients with firearm injuries. Methods/approach We identified all firearm injuries in the National Trauma Data Bank, 2009–2012. Analyses were stratified by age group (0–17/18–64/65–89) and treatment setting (ED/inpatient). The 6-hour (ED) and 30-day (inpatient) discharge survival by injury intent were estimated using the Kaplan-Meier Method. Clustered multivariate Poisson regression models were used to calculate risk ratios (RR) of mortality in each setting. Results Records of 67,212 firearm injuries were analysed. ED patients were demographically similar to inpatients. Mortality rates in the ED were higher than in the inpatient setting across intent categories and age groups. Intent was strongly associated with mortality, with evidence of race modifying this association. In the ED setting, risk of mortality was higher with self-inflicted injuries (RR = 6.65; 95% confidence interval (CI): 4.52, 9.77), assault injuries (RR = 2.01; 95% CI: 1.34, 3.01), and undetermined intent injuries (RR = 4.45; 95% CI: 2.06, 9.61) compared to unintentional injuries. In the inpatient setting, risk of mortality was higher with self-inflicted injuries (RR = 8.28; 95% CI: 7.10, 9.67), assault injuries (RR = 2.02; 95% CI: 1.73, 2.36), and undetermined intent injuries (RR = 3.04; 95% CI: 2.42, 3.08) compared to unintentional injuries. Conclusions Patterns of firearm injury intent differ between ED and inpatient settings. The risk of death in ED and inpatient settings is notably higher for intentional than unintentional injuries. Significance and contributions Accounting for injury intent identifies subgroups at particular risk of firearm mortality in different treatment settings.