Predicting Intensive Care Delirium with Machine Learning: Model Development and External Validation.

BACKGROUND: Delirium poses significant risks to patients, but countermeasures can be taken to mitigate negative outcomes. Accurately forecasting delirium in intensive care unit (ICU) patients could guide proactive intervention. Our primary objective was to predict ICU delirium by applying machine learning to clinical and physiologic data routinely collected in electronic health records. METHODS: Two prediction models were trained and tested using a multicenter database (years of data collection 2014 to 2015), and externally validated on two single-center databases (2001 to 2012 and 2008 to 2019). The primary outcome variable was delirium defined as a positive Confusion Assessment Method for the ICU screen, or an Intensive Care Delirium Screening Checklist of 4 or greater. The first model, named "24-hour model," used data from the 24 h after ICU admission to predict delirium any time afterward. The second model designated "dynamic model," predicted the onset of delirium up to 12 h in advance. Model performance was compared with a widely cited reference model. RESULTS: For the 24-h model, delirium was identified in 2,536 of 18,305 (13.9%), 768 of 5,299 (14.5%), and 5,955 of 36,194 (11.9%) of patient stays, respectively, in the development sample and two validation samples. For the 12-h lead time dynamic model, delirium was identified in 3,791 of 22,234 (17.0%), 994 of 6,166 (16.1%), and 5,955 of 28,440 (20.9%) patient stays, respectively. Mean area under the receiver operating characteristics curve (AUC) (95% CI) for the first 24-h model was 0.785 (0.769 to 0.801), significantly higher than the modified reference model with AUC of 0.730 (0.704 to 0.757). The dynamic model had a mean AUC of 0.845 (0.831 to 0.859) when predicting delirium 12 h in advance. Calibration was similar in both models (mean Brier Score [95% CI] 0.102 [0.097 to 0.108] and 0.111 [0.106 to 0.116]). Model discrimination and calibration were maintained when tested on the validation datasets. CONCLUSIONS: Machine learning models trained with routinely collected electronic health record data accurately predict ICU delirium, supporting dynamic time-sensitive forecasting.

Discovery of novel cMET-targeting antibody Fab drug conjugates as potential treatment for solid tumors with highly expressed cMET.

BACKGROUND: Solid tumors are becoming prevalent affecting both old and young populations. Numerous solid tumors are associated with high cMET expression. The complexity of solid tumors combined with the highly interconnected nature of the cMET/HGF pathway with other cellular pathways make the pursuit of finding an effective treatment extremely challenging. The current standard of care for these malignancies is mostly small molecule-based chemotherapy. Antibody-based therapeutics as well as antibody drug conjugates are promising emerging classes against cMET-overexpressing solid tumors. RESEARCH DESIGN AND METHODS: In this study, we described the design, synthesis, in vitro and in vivo characterization of cMET-targeting Fab drug conjugates (FDCs) as an alternative therapeutic strategy. The format is comprised of a Fab conjugated to a potent cytotoxic drug via a cleavable linker employing lysine-based and cysteine-based conjugation chemistries. RESULTS: We found that the FDCs have potent anti-tumor efficacies in cancer cells with elevated overexpression of cMET. Moreover, they demonstrated a remarkable anti-tumor effect in a human gastric xenograft mouse model. CONCLUSIONS: The FDC format has the potential to overcome some of the challenges presented by the other classes of therapeutics. This study highlights the promise of antibody fragment-based drug conjugate formats for the treatment of solid tumors.

The association of Medicaid expansion and racial/ethnic inequities in access, treatment, and outcomes for patients with acute myocardial infarction.

INTRODUCTION: After having an acute myocardial infarction (AMI), racial and ethnic minorities have less access to care, decreased rates of invasive treatments such as percutaneous coronary intervention (PCI), and worse outcomes compared with white patients. The objective of this study was to determine whether the Affordable Care Act's expansion of Medicaid eligibility was associated with changes in racial disparities in access, treatments, and outcomes after AMI. METHODS: Quasi-experimental, difference-in-differences-in-differences analysis of non-Hispanic white and minority patients with acute myocardial infarction in California and Florida from 2010-2015, using linear regression models to estimate the difference-in-differences. This population-based sample included all Medicaid and uninsured patients ages 18-64 hospitalized with acute myocardial infarction in California, which expanded Medicaid through the Affordable Care Act beginning as early as July 2011 in certain counties, and Florida, which did not expand Medicaid. The main outcomes included rates of admission to hospitals capable of performing PCI, rates of transfer for patients who first presented to hospitals that did not perform PCI, rates of PCI during hospitalization and rates of early (within 48 hours of admission) PCI, rates of readmission to the hospital within 30 days, and rates of in-hospital mortality. RESULTS: A total of 55,991 hospital admissions met inclusion criteria, 32,540 of which were in California and 23,451 were in Florida. Among patients with AMI who initially presented to a non-PCI hospital, the likelihood of being transferred increased by 12 percentage points (95% CI 2 to 21) for minority patients relative to white patients after the Medicaid expansion. The likelihood of undergoing PCI increased by 3 percentage points (95% CI 0 to 5) for minority patients relative to white patients after the Medicaid expansion. We did not find an association between the Medicaid expansion and racial disparities in overall likelihood of admission to a PCI hospital, hospital readmissions, or in-hospital mortality. CONCLUSIONS: The Medicaid expansion was associated with a decrease in racial disparities in transfers and rates of PCI after AMI. We did not find an association between the Medicaid expansion and admission to a PCI hospital, readmissions, and in-hospital mortality. Additional factors outside of insurance coverage likely continue to contribute to disparities in outcomes after AMI. These findings are crucial for policy makers seeking to reduce racial disparities in access, treatment and outcomes in AMI.

Effects of Medicaid expansion on access, treatment and outcomes for patients with acute myocardial infarction.

INTRODUCTION: Uninsured patients have decreased access to care, lower rates of percutaneous coronary intervention (PCI), and worse outcomes after acute myocardial infarction (AMI). The aim of this study was to determine whether expanding insurance coverage through the Affordable Care Act's expansion of Medicaid eligibility affected access to PCI hospitals, rates of PCI, 30-day readmissions, and in-hospital mortality after AMI. METHODS: Quasi-experimental, difference-in-differences analysis of Medicaid and uninsured patients with acute myocardial infarction in California, which expanded Medicaid through the Affordable Care Act, and Florida, which did not, from 2010-2015. This study accounts for the early expansion of Medicaid in certain California counties that began as early as July 2011. Main outcomes included rates of admission to PCI hospitals, rates of transfer for patients who initially presented to non-PCI hospitals, rates of PCI, rates of early PCI defined as within 48 hours of hospital admission, in-hospital mortality, and 30-day readmission. RESULTS: 55,991 hospital admissions between 2010-2015 met inclusion criteria. Of these, 32,540 were in California, which expanded Medicaid, and 23,451 were in Florida, which did not. 30-day readmission rates after AMI decreased by an absolute difference of 1.22 percentage points after the Medicaid expansion (95% CI -2.14 to -0.30, P < 0.01). This represented a relative decrease in readmission rates of 9.5% after AMI. No relationship between the Medicaid expansion and admission to PCI hospitals, transfer to PCI hospitals, rates of PCI, rates of early PCI, or in-hospital mortality were observed. CONCLUSIONS: Hospital readmissions decreased by 9.5% after the Affordable Care Act expanded Medicaid eligibility, although there was no association found between Medicaid expansion and access to PCI hospitals or treatment with PCI. Better understanding the ways that Medicaid expansion might affect care for vulnerable populations with AMI is important for policymakers considering whether to expand Medicaid eligibility in their state.

Revisits, readmissions, and outcomes for pediatric traumatic brain injury in California, 2005-2014.

Long-term outcomes related to emergency department revisit, hospital readmission, and all-cause mortality, have not been well characterized across the spectrum of pediatric traumatic brain injury (TBI). We evaluated emergency department visit outcomes up to 1 year after pediatric TBI, in comparison to a referent group of trauma patients without TBI. We performed a longitudinal, retrospective study of all pediatric trauma patients who presented to emergency departments and hospitals in California from 2005 to 2014. We compared emergency department visits, dispositions, revisits, readmissions, and mortality in pediatric trauma patients with a TBI diagnosis to those without TBI (Other Trauma patients). We identified 208,222 pediatric patients with an index diagnosis of TBI and 1,314,064 patients with an index diagnosis of Other Trauma. Population growth adjusted TBI visits increased by 5.6% while those for Other Trauma decreased by 40.7%. The majority of patients were discharged from the emergency department on their first visit (93.2% for traumatic brain injury vs. 96.5% for Other Trauma). A greater proportion of TBI patients revisited the emergency department (33.4% vs. 3.0%) or were readmitted to the hospital (0.9% vs. 0.04%) at least once within a year of discharge. The health burden within a year after a pediatric TBI visit is considerable and is greater than that of non-TBI trauma. These data suggest that outpatient strategies to monitor for short-term and longer-term sequelae after pediatric TBI are needed to improve patient outcomes, lessen the burden on families, and more appropriately allocate resources in the healthcare system.

Ten-year trends in traumatic brain injury: a retrospective cohort study of California emergency department and hospital revisits and readmissions.

OBJECTIVE: To describe visits and visit rates of adults presenting to emergency departments (EDs) with a diagnosis of traumatic brain injury (TBI). TBI is a major cause of death and disability in the USA; yet, current literature is limited because few studies examine longer-term ED revisits and hospital readmission patterns of TBI patients across a broad spectrum of injury severity, which can help inform potential unmet healthcare needs. DESIGN: We performed a retrospective cohort study. SETTING: We analysed non-public patient-level data from California's Office of Statewide Health Planning and Development for years 2005 to 2014. PARTICIPANTS: We identified 1.2 million adult patients aged ≥18 years presenting to California EDs and hospitals with an index diagnosis of TBI. PRIMARY AND SECONDARY OUTCOME MEASURES: Our main outcomes included revisits, readmissions and mortality over time. We also examined demographics, mechanism and severity of injury and disposition at discharge. RESULTS: We found a 57.7% increase in the number of TBI ED visits, representing a 40.5% increase in TBI visit rates over the 10-year period (346-487 per 100 000 residents). During this time, there was also a 33.8% decrease in the proportion of patients admitted to the hospital. Older, publicly insured and black populations had the highest visit rates, and falls were the most common mechanism of injury (45.5% of visits). Of all patients with an index TBI visit, 40.5% of them had a revisit during the first year, with 46.7% of them seeking care at a different hospital from their initial hospital or ED visit. Additionally, of revisits within the first year, 13.4% of them resulted in hospital readmission. CONCLUSIONS: The large proportion of patients with TBI who are discharged directly from the ED, along with the high rates of revisits and readmissions, suggest a role for an established system for follow-up, treatment and care of TBI.

Trends in the utilisation of emergency departments in California, 2005-2015: a retrospective analysis.

OBJECTIVE: To examine current trends in the characteristics of patients visiting California emergency departments (EDs) in order to better direct the allocation of acute care resources. DESIGN: A retrospective study. SETTING: We analysed ED utilisation trends between 2005 and 2015 in California using non-public patient data from California's Office of Statewide Health Planning and Development. PARTICIPANTS: We included all ED visits in California from 2005 to 2015. PRIMARY AND SECONDARY OUTCOME MEASURES: We analysed ED visits and visit rates by age, sex, race/ethnicity, payer and urban/rural trends. We further examined age, sex, race/ethnicity and urban/rural trends within each payer group for a more granular picture of the patient population. Additionally, we looked at the proportion of patients admitted from the ED and distribution of diagnoses. RESULTS: Between 2005 and 2015, the annual number of ED visits increased from 10.2 to 14.2 million in California. ED visit rates increased by 27.8% (p<0.001), with the greatest increases among patients aged 5-19 (37.4%, p<0.001) and 45-64 years (41.1%, p<0.001), non-Hispanic Black and Hispanic patients (56.8% and 48.8%, p<0.001), the uninsured and Medicaid-insured (36.1%, p=0.002; 28.6%, p<0.001) and urban residents (28.3%, p<0.001). The proportion of ED visits resulting in hospitalisation decreased by 18.3%, with decreases across all payer groups. CONCLUSIONS: Our findings reveal an increasing demand for emergency care and may reflect current limitations in accessing care in other parts of the healthcare system. Policymakers may need to recognise the increasingly vital role that EDs are playing in the provision of care and consider ways to incorporate this changing reality into the delivery of health services.

Erratum: This Article Corrects: "Trends in Regionalization of Care for ST-Segment Elevation Myocardial Infarction".

[This corrects the article on p. 1010 in vol. 18, PMID: 29085531.].

Trends in Regionalization of Care for ST-Segment Elevation Myocardial Infarction.

INTRODUCTION: California has led successful regionalized efforts for several time-critical medical conditions, including ST-segment elevation myocardial infarction (STEMI), but no specific mandated protocols exist to define regionalization of care. We aimed to study the trends in regionalization of care for STEMI patients in the state of California and to examine the differences in patient demographic, hospital, and county trends. METHODS: Using survey responses collected from all California emergency medical services (EMS) agencies, we developed four categories - no, partial, substantial, and complete regionalization - to capture prehospital and inter-hospital components of regionalization in each EMS agency's jurisdiction between 2005-2014. We linked the survey responses to 2006 California non-public hospital discharge data to study the patient distribution at baseline. RESULTS: STEMI regionalization-of-care networks steadily developed across California. Only 14% of counties were regionalized in 2006, accounting for 42% of California's STEMI patient population, but over half of these counties, representing 86% of California's STEMI patient population, reached complete regionalization in 2014. We did not find any dramatic differences in underlying patient characteristics based on regionalization status; however, differences in hospital characteristics were relatively substantial. CONCLUSION: Potential barriers to achieving regionalization included competition, hospital ownership, population density, and financial challenges. Minimal differences in patient characteristics can establish that patient differences unlikely played any role in influencing earlier or later regionalization and can provide a framework for future analyses evaluating the impact of regionalization on patient outcomes.