Measuring quality-of-care in treatment of young children with attention-deficit/hyperactivity disorder using pre-trained language models

OBJECTIVE

To measure pediatrician adherence to evidence-based guidelines in the treatment of young children with attention-deficit/hyperactivity disorder (ADHD) in a diverse healthcare system using natural language processing (NLP) techniques.

MATERIALS AND METHODS

We extracted structured and free-text data from electronic health records (EHRs) of all office visits (2015-2019) of children aged 4-6 years in a community-based primary healthcare network in California, who had ≥1 visits with an ICD-10 diagnosis of ADHD. Two pediatricians annotated clinical notes of the first ADHD visit for 423 patients. Inter-annotator agreement (IAA) was assessed for the recommendation for the first-line behavioral treatment (F-measure = 0.89). Four pre-trained language models, including BioClinical Bidirectional Encoder Representations from Transformers (BioClinicalBERT), were used to identify behavioral treatment recommendations using a 70/30 train/test split. For temporal validation, we deployed BioClinicalBERT on 1,020 unannotated notes from other ADHD visits and well-care visits; all positively classified notes (n = 53) and 5% of negatively classified notes (n = 50) were manually reviewed.

RESULTS

Of 423 patients, 313 (74%) were male; 298 (70%) were privately insured; 138 (33%) were White; 61 (14%) were Hispanic. The BioClinicalBERT model trained on the first ADHD visits achieved F1 = 0.76, precision = 0.81, recall = 0.72, and AUC = 0.81 [0.72-0.89]. Temporal validation achieved F1 = 0.77, precision = 0.68, and recall = 0.88. Fairness analysis revealed low model performance in publicly insured patients (F1 = 0.53).

CONCLUSION

Deploying pre-trained language models on a variable set of clinical notes accurately captured pediatrician adherence to guidelines in the treatment of children with ADHD. Validating this approach in other patient populations is needed to achieve equitable measurement of quality of care at scale and improve clinical care for mental health conditions.

Characterizing the limitations of using diagnosis codes in the context of machine learning for healthcare

BACKGROUND

Diagnostic codes are commonly used as inputs for clinical prediction models, to create labels for prediction tasks, and to identify cohorts for multicenter network studies. However, the coverage rates of diagnostic codes and their variability across institutions are underexplored. The primary objective was to describe lab- and diagnosis-based labels for 7 selected outcomes at three institutions. Secondary objectives were to describe agreement, sensitivity, and specificity of diagnosis-based labels against lab-based labels.

METHODS

This study included three cohorts: SickKids from The Hospital for Sick Children, and StanfordPeds and StanfordAdults from Stanford Medicine. We included seven clinical outcomes with lab-based definitions: acute kidney injury, hyperkalemia, hypoglycemia, hyponatremia, anemia, neutropenia and thrombocytopenia. For each outcome, we created four lab-based labels (abnormal, mild, moderate and severe) based on test result and one diagnosis-based label. Proportion of admissions with a positive label were presented for each outcome stratified by cohort. Using lab-based labels as the gold standard, agreement using Cohen's Kappa, sensitivity and specificity were calculated for each lab-based severity level.

RESULTS

The number of admissions included were: SickKids (n = 59,298), StanfordPeds (n = 24,639) and StanfordAdults (n = 159,985). The proportion of admissions with a positive diagnosis-based label was significantly higher for StanfordPeds compared to SickKids across all outcomes, with odds ratio (99.9% confidence interval) for abnormal diagnosis-based label ranging from 2.2 (1.7-2.7) for neutropenia to 18.4 (10.1-33.4) for hyperkalemia. Lab-based labels were more similar by institution. When using lab-based labels as the gold standard, Cohen's Kappa and sensitivity were lower at SickKids for all severity levels compared to StanfordPeds.

CONCLUSIONS

Across multiple outcomes, diagnosis codes were consistently different between the two pediatric institutions. This difference was not explained by differences in test results. These results may have implications for machine learning model development and deployment.

Predicting patients who are likely to develop Lupus Nephritis of those newly diagnosed with Systemic Lupus Erythematosus

Patients diagnosed with systemic lupus erythematosus (SLE) suffer from a decreased quality of life, an increased risk of medical complications, and an increased risk of death. In particular, approximately 50% of SLE patients progress to develop lupus nephritis, which oftentimes leads to life-threatening end stage renal disease (ESRD) and requires dialysis or kidney transplant1. The challenge is that lupus nephritis is diagnosed via a kidney biopsy, which is typically performed only after noticeable decreased kidney function, leaving little room for proactive or preventative measures. The ability to predict which patients are most likely to develop lupus nephritis has the potential to shift lupus nephritis disease management from reactive to proactive. We present a clinically useful prediction model to predict which patients with newly diagnosed SLE will go on to develop lupus nephritis in the next five years.

EHR foundation models improve robustness in the presence of temporal distribution shift

Temporal distribution shift negatively impacts the performance of clinical prediction models over time. Pretraining foundation models using self-supervised learning on electronic health records (EHR) may be effective in acquiring informative global patterns that can improve the robustness of task-specific models. The objective was to evaluate the utility of EHR foundation models in improving the in-distribution (ID) and out-of-distribution (OOD) performance of clinical prediction models. Transformer- and gated recurrent unit-based foundation models were pretrained on EHR of up to 1.8 M patients (382 M coded events) collected within pre-determined year groups (e.g., 2009-2012) and were subsequently used to construct patient representations for patients admitted to inpatient units. These representations were used to train logistic regression models to predict hospital mortality, long length of stay, 30-day readmission, and ICU admission. We compared our EHR foundation models with baseline logistic regression models learned on count-based representations (count-LR) in ID and OOD year groups. Performance was measured using area-under-the-receiver-operating-characteristic curve (AUROC), area-under-the-precision-recall curve, and absolute calibration error. Both transformer and recurrent-based foundation models generally showed better ID and OOD discrimination relative to count-LR and often exhibited less decay in tasks where there is observable degradation of discrimination performance (average AUROC decay of 3% for transformer-based foundation model vs. 7% for count-LR after 5-9 years). In addition, the performance and robustness of transformer-based foundation models continued to improve as pretraining set size increased. These results suggest that pretraining EHR foundation models at scale is a useful approach for developing clinical prediction models that perform well in the presence of temporal distribution shift.

Evaluation of Feature Selection Methods for Preserving Machine Learning Performance in the Presence of Temporal Dataset Shift in Clinical Medicine

BACKGROUND

 Temporal dataset shift can cause degradation in model performance as discrepancies between training and deployment data grow over time. The primary objective was to determine whether parsimonious models produced by specific feature selection methods are more robust to temporal dataset shift as measured by out-of-distribution (OOD) performance, while maintaining in-distribution (ID) performance.

METHODS

 Our dataset consisted of intensive care unit patients from MIMIC-IV categorized by year groups (2008-2010, 2011-2013, 2014-2016, and 2017-2019). We trained baseline models using L2-regularized logistic regression on 2008-2010 to predict in-hospital mortality, long length of stay (LOS), sepsis, and invasive ventilation in all year groups. We evaluated three feature selection methods: L1-regularized logistic regression (L1), Remove and Retrain (ROAR), and causal feature selection. We assessed whether a feature selection method could maintain ID performance (2008-2010) and improve OOD performance (2017-2019). We also assessed whether parsimonious models retrained on OOD data performed as well as oracle models trained on all features in the OOD year group.

RESULTS

 The baseline model showed significantly worse OOD performance with the long LOS and sepsis tasks when compared with the ID performance. L1 and ROAR retained 3.7 to 12.6% of all features, whereas causal feature selection generally retained fewer features. Models produced by L1 and ROAR exhibited similar ID and OOD performance as the baseline models. The retraining of these models on 2017-2019 data using features selected from training on 2008-2010 data generally reached parity with oracle models trained directly on 2017-2019 data using all available features. Causal feature selection led to heterogeneous results with the superset maintaining ID performance while improving OOD calibration only on the long LOS task.

CONCLUSIONS

 While model retraining can mitigate the impact of temporal dataset shift on parsimonious models produced by L1 and ROAR, new methods are required to proactively improve temporal robustness.

Characteristics and outcomes of COVID-19 patients with COPD from the United States, South Korea, and Europe

Background: Characterization studies of COVID-19 patients with chronic obstructive pulmonary disease (COPD) are limited in size and scope. The aim of the study is to provide a large-scale characterization of COVID-19 patients with COPD. Methods: We included thirteen databases contributing data from January-June 2020 from North America (US), Europe and Asia. We defined two cohorts of patients with COVID-19 namely a 'diagnosed' and 'hospitalized' cohort. We followed patients from COVID-19 index date to 30 days or death. We performed descriptive analysis and reported the frequency of characteristics and outcomes among COPD patients with COVID-19. Results: The study included 934,778 patients in the diagnosed COVID-19 cohort and 177,201 in the hospitalized COVID-19 cohort. Observed COPD prevalence in the diagnosed cohort ranged from 3.8% (95%CI 3.5-4.1%) in French data to 22.7% (95%CI 22.4-23.0) in US data, and from 1.9% (95%CI 1.6-2.2) in South Korean to 44.0% (95%CI 43.1-45.0) in US data, in the hospitalized cohorts. COPD patients in the hospitalized cohort had greater comorbidity than those in the diagnosed cohort, including hypertension, heart disease, diabetes and obesity. Mortality was higher in COPD patients in the hospitalized cohort and ranged from 7.6% (95%CI 6.9-8.4) to 32.2% (95%CI 28.0-36.7) across databases. ARDS, acute renal failure, cardiac arrhythmia and sepsis were the most common outcomes among hospitalized COPD patients.   Conclusion: COPD patients with COVID-19 have high levels of COVID-19-associated comorbidities and poor COVID-19 outcomes. Further research is required to identify patients with COPD at high risk of worse outcomes.

Perspective Toward Machine Learning Implementation in Pediatric Medicine: Mixed Methods Study

Background

Given the costs of machine learning implementation, a systematic approach to prioritizing which models to implement into clinical practice may be valuable.

Objective

The primary objective was to determine the health care attributes respondents at 2 pediatric institutions rate as important when prioritizing machine learning model implementation. The secondary objective was to describe their perspectives on implementation using a qualitative approach.

Methods

In this mixed methods study, we distributed a survey to health system leaders, physicians, and data scientists at 2 pediatric institutions. We asked respondents to rank the following 5 attributes in terms of implementation usefulness: the clinical problem was common, the clinical problem caused substantial morbidity and mortality, risk stratification led to different actions that could reasonably improve patient outcomes, reducing physician workload, and saving money. Important attributes were those ranked as first or second most important. Individual qualitative interviews were conducted with a subsample of respondents.

Results

Among 613 eligible respondents, 275 (44.9%) responded. Qualitative interviews were conducted with 17 respondents. The most common important attributes were risk stratification leading to different actions (205/275, 74.5%) and clinical problem causing substantial morbidity or mortality (177/275, 64.4%). The attributes considered least important were reducing physician workload and saving money. Qualitative interviews consistently prioritized implementations that improved patient outcomes.

Conclusions

Respondents prioritized machine learning model implementation where risk stratification would lead to different actions and clinical problems that caused substantial morbidity and mortality. Implementations that improved patient outcomes were prioritized. These results can help provide a framework for machine learning model implementation.

Characteristics and outcomes of COVID-19 patients with COPD from the United States, South Korea, and Europe

Background: Characterization studies of COVID-19 patients with chronic obstructive pulmonary disease (COPD) are limited in size and scope. The aim of the study is to provide a large-scale characterization of COVID-19 patients with COPD. Methods: We included thirteen databases contributing data from January-June 2020 from North America (US), Europe and Asia. We defined two cohorts of patients with COVID-19 namely a 'diagnosed' and 'hospitalized' cohort. We followed patients from COVID-19 index date to 30 days or death. We performed descriptive analysis and reported the frequency of characteristics and outcomes among COPD patients with COVID-19. Results: The study included 934,778 patients in the diagnosed COVID-19 cohort and 177,201 in the hospitalized COVID-19 cohort. Observed COPD prevalence in the diagnosed cohort ranged from 3.8% (95%CI 3.5-4.1%) in French data to 22.7% (95%CI 22.4-23.0) in US data, and from 1.9% (95%CI 1.6-2.2) in South Korean to 44.0% (95%CI 43.1-45.0) in US data, in the hospitalized cohorts. COPD patients in the hospitalized cohort had greater comorbidity than those in the diagnosed cohort, including hypertension, heart disease, diabetes and obesity. Mortality was higher in COPD patients in the hospitalized cohort and ranged from 7.6% (95%CI 6.9-8.4) to 32.2% (95%CI 28.0-36.7) across databases. ARDS, acute renal failure, cardiac arrhythmia and sepsis were the most common outcomes among hospitalized COPD patients.   Conclusion: COPD patients with COVID-19 have high levels of COVID-19-associated comorbidities and poor COVID-19 outcomes. Further research is required to identify patients with COPD at high risk of worse outcomes.

Building a Learning Health System: Creating an Analytical Workflow for Evidence Generation to Inform Institutional Clinical Care Guidelines

Background  One key aspect of a learning health system (LHS) is utilizing data generated during care delivery to inform clinical care. However, institutional guidelines that utilize observational data are rare and require months to create, making current processes impractical for more urgent scenarios such as those posed by the COVID-19 pandemic. There exists a need to rapidly analyze institutional data to drive guideline creation where evidence from randomized control trials are unavailable.

Objectives  This article provides a background on the current state of observational data generation in institutional guideline creation and details our institution's experience in creating a novel workflow to (1) demonstrate the value of such a workflow, (2) demonstrate a real-world example, and (3) discuss difficulties encountered and future directions.

Methods  Utilizing a multidisciplinary team of database specialists, clinicians, and informaticists, we created a workflow for identifying and translating a clinical need into a queryable format in our clinical data warehouse, creating data summaries and feeding this information back into clinical guideline creation.

Results  Clinical questions posed by the hospital medicine division were answered in a rapid time frame and informed creation of institutional guidelines for the care of patients with COVID-19. The cost of setting up a workflow, answering the questions, and producing data summaries required around 300 hours of effort and $300,000 USD.

Conclusion  A key component of an LHS is the ability to learn from data generated during care delivery. There are rare examples in the literature and we demonstrate one such example along with proposed thoughts of ideal multidisciplinary team formation and deployment.

Evaluation of domain generalization and adaptation on improving model robustness to temporal dataset shift in clinical medicine

Temporal dataset shift associated with changes in healthcare over time is a barrier to deploying machine learning-based clinical decision support systems. Algorithms that learn robust models by estimating invariant properties across time periods for domain generalization (DG) and unsupervised domain adaptation (UDA) might be suitable to proactively mitigate dataset shift. The objective was to characterize the impact of temporal dataset shift on clinical prediction models and benchmark DG and UDA algorithms on improving model robustness. In this cohort study, intensive care unit patients from the MIMIC-IV database were categorized by year groups (2008–2010, 2011–2013, 2014–2016 and 2017–2019). Tasks were predicting mortality, long length of stay, sepsis and invasive ventilation. Feedforward neural networks were used as prediction models. The baseline experiment trained models using empirical risk minimization (ERM) on 2008–2010 (ERM[08–10]) and evaluated them on subsequent year groups. DG experiment trained models using algorithms that estimated invariant properties using 2008–2016 and evaluated them on 2017–2019. UDA experiment leveraged unlabelled samples from 2017 to 2019 for unsupervised distribution matching. DG and UDA models were compared to ERM[08–16] models trained using 2008–2016. Main performance measures were area-under-the-receiver-operating-characteristic curve (AUROC), area-under-the-precision-recall curve and absolute calibration error. Threshold-based metrics including false-positives and false-negatives were used to assess the clinical impact of temporal dataset shift and its mitigation strategies. In the baseline experiments, dataset shift was most evident for sepsis prediction (maximum AUROC drop, 0.090; 95% confidence interval (CI), 0.080–0.101). Considering a scenario of 100 consecutively admitted patients showed that ERM[08–10] applied to 2017–2019 was associated with one additional false-negative among 11 patients with sepsis, when compared to the model applied to 2008–2010. When compared with ERM[08–16], DG and UDA experiments failed to produce more robust models (range of AUROC difference, − 0.003 to 0.050). In conclusion, DG and UDA failed to produce more robust models compared to ERM in the setting of temporal dataset shift. Alternate approaches are required to preserve model performance over time in clinical medicine.

Characteristics and outcomes of COVID-19 patients with COPD from the United States, South Korea, and Europe

Background: Characterization studies of COVID-19 patients with chronic obstructive pulmonary disease (COPD) are limited in size and scope. The aim of the study is to provide a large-scale characterization of COVID-19 patients with COPD. Methods: We included thirteen databases contributing data from January-June 2020 from North America (US), Europe and Asia. We defined two cohorts of patients with COVID-19 namely a 'diagnosed' and 'hospitalized' cohort. We followed patients from COVID-19 index date to 30 days or death. We performed descriptive analysis and reported the frequency of characteristics and outcomes among COPD patients with COVID-19. Results: The study included 934,778 patients in the diagnosed COVID-19 cohort and 177,201 in the hospitalized COVID-19 cohort. Observed COPD prevalence in the diagnosed cohort ranged from 3.8% (95%CI 3.5-4.1%) in French data to 22.7% (95%CI 22.4-23.0) in US data, and from 1.9% (95%CI 1.6-2.2) in South Korean to 44.0% (95%CI 43.1-45.0) in US data, in the hospitalized cohorts. COPD patients in the hospitalized cohort had greater comorbidity than those in the diagnosed cohort, including hypertension, heart disease, diabetes and obesity. Mortality was higher in COPD patients in the hospitalized cohort and ranged from 7.6% (95%CI 6.9-8.4) to 32.2% (95%CI 28.0-36.7) across databases. ARDS, acute renal failure, cardiac arrhythmia and sepsis were the most common outcomes among hospitalized COPD patients.   Conclusion: COPD patients with COVID-19 have high levels of COVID-19-associated comorbidities and poor COVID-19 outcomes. Further research is required to identify patients with COPD at high risk of worse outcomes.

Rate of Pediatrician Recommendations for Behavioral Treatment for Preschoolers With Attention-Deficit/Hyperactivity Disorder Diagnosis or Related Symptoms

This cohort study investigates the rate of pediatrician recommendations for behavioral treatment for preschoolers with an attention-deficit/hyperactivity disorder diagnosis or symptoms.

Systematic Review of Approaches to Preserve Machine Learning Performance in the Presence of Temporal Dataset Shift in Clinical Medicine

OBJECTIVE

The change in performance of machine learning models over time as a result of temporal dataset shift is a barrier to machine learning-derived models facilitating decision-making in clinical practice. Our aim was to describe technical procedures used to preserve the performance of machine learning models in the presence of temporal dataset shifts.

METHODS

Studies were included if they were fully published articles that used machine learning and implemented a procedure to mitigate the effects of temporal dataset shift in a clinical setting. We described how dataset shift was measured, the procedures used to preserve model performance, and their effects.

RESULTS

Of 4,457 potentially relevant publications identified, 15 were included. The impact of temporal dataset shift was primarily quantified using changes, usually deterioration, in calibration or discrimination. Calibration deterioration was more common (n = 11) than discrimination deterioration (n = 3). Mitigation strategies were categorized as model level or feature level. Model-level approaches (n = 15) were more common than feature-level approaches (n = 2), with the most common approaches being model refitting (n = 12), probability calibration (n = 7), model updating (n = 6), and model selection (n = 6). In general, all mitigation strategies were successful at preserving calibration but not uniformly successful in preserving discrimination.

CONCLUSION

There was limited research in preserving the performance of machine learning models in the presence of temporal dataset shift in clinical medicine. Future research could focus on the impact of dataset shift on clinical decision making, benchmark the mitigation strategies on a wider range of datasets and tasks, and identify optimal strategies for specific settings.

Ontology-driven weak supervision for clinical entity classification in electronic health records

In the electronic health record, using clinical notes to identify entities such as disorders and their temporality (e.g. the order of an event relative to a time index) can inform many important analyses. However, creating training data for clinical entity tasks is time consuming and sharing labeled data is challenging due to privacy concerns. The information needs of the COVID-19 pandemic highlight the need for agile methods of training machine learning models for clinical notes. We present Trove, a framework for weakly supervised entity classification using medical ontologies and expert-generated rules. Our approach, unlike hand-labeled notes, is easy to share and modify, while offering performance comparable to learning from manually labeled training data. In this work, we validate our framework on six benchmark tasks and demonstrate Trove's ability to analyze the records of patients visiting the emergency department at Stanford Health Care for COVID-19 presenting symptoms and risk factors.

Unraveling COVID-19: a large-scale characterization of 4.5 million COVID-19 cases using CHARYBDIS

Background: Routinely collected real world data (RWD) have great utility in aiding the novel coronavirus disease (COVID-19) pandemic response [1,2]. Here we present the international Observational Health Data Sciences and Informatics (OHDSI) [3] Characterizing Health Associated Risks, and Your Baseline Disease In SARS-COV-2 (CHARYBDIS) framework for standardisation and analysis of COVID-19 RWD. Methods: We conducted a descriptive cohort study using a federated network of data partners in the United States, Europe (the Netherlands, Spain, the UK, Germany, France and Italy) and Asia (South Korea and China). The study protocol and analytical package were released on 11 th June 2020 and are iteratively updated via GitHub [4]. Findings: We identified three non-mutually exclusive cohorts of 4,537,153 individuals with a clinical COVID-19 diagnosis or positive test, 886,193 hospitalized with COVID-19 , and 113,627 hospitalized with COVID-19 requiring intensive services . All comorbidities, symptoms, medications, and outcomes are described by cohort in aggregate counts, and are available in an interactive website: https://data.ohdsi.org/Covid19CharacterizationCharybdis/. Interpretation: CHARYBDIS findings provide benchmarks that contribute to our understanding of COVID-19 progression, management and evolution over time. This can enable timely assessment of real-world outcomes of preventative and therapeutic options as they are introduced in clinical practice.

Characterizing database granularity using SNOMED-CT hierarchy

Multi-center observational studies require recognition and reconciliation of differences in patient representations arising from underlying populations, disparate coding practices and specifics of data capture. This leads to different granularity or detail of concepts representing the clinical facts. For researchers studying certain populations of interest, it is important to ensure that concepts at the right level are used for the definition of these populations. We studied the granularity of concepts within 22 data sources in the OHDSI network and calculated a composite granularity score for each dataset. Three alternative SNOMED-based approaches for such score showed consistency in classifying data sources into three levels of granularity (low, moderate and high), which correlated with the provenance of data and country of origin. However, they performed unsatisfactorily in ordering data sources within these groups and showed inconsistency for small data sources. Further studies on examining approaches to data source granularity are needed.

Ontology-driven weak supervision for clinical entity classification in electronic health records

In the electronic health record, using clinical notes to identify entities such as disorders and their temporality (e.g. the order of an event relative to a time index) can inform many important analyses. However, creating training data for clinical entity tasks is time consuming and sharing labeled data is challenging due to privacy concerns. The information needs of the COVID-19 pandemic highlight the need for agile methods of training machine learning models for clinical notes. We present Trove, a framework for weakly supervised entity classification using medical ontologies and expert-generated rules. Our approach, unlike hand-labeled notes, is easy to share and modify, while offering performance comparable to learning from manually labeled training data. In this work, we validate our framework on six benchmark tasks and demonstrate Trove's ability to analyze the records of patients visiting the emergency department at Stanford Health Care for COVID-19 presenting symptoms and risk factors.

A Guide for Pain Management in Low and Middle Income Communities. Managing the Risk of Opioid Abuse in Patients with Cancer Pain

Most patients who present with cancer have advanced disease and often suffer moderate to severe pain. Opioid therapy can be safe and effective for use in cancer patients with pain, but there are rightful concerns about inappropriate opioid use even in the cancer population. Since cancer patients live longer than ever before in history (and survivors may have long exposure times to opioid therapy), opioid misuse among cancer patients is an important topic worthy of deeper investigation. Cancer patients with pain must be evaluated for risk factors for potential opioid misuse and aberrant drug-taking behaviors assessed. A variety of validated screening tools should be used. Of particular importance is the fact that pain in cancer patients changes frequently, whether it is related to their underlying disease (progression or remission), pain related to treatment (such as painful chemotherapy-induced peripheral neuropathy), and concomitant pain unrelated to cancer (such as osteoarthritis, headache, or back pain). Fortunately, clinicians can use universal precautions to help reduce the risk of opioid misuse while still assuring that cancer patients get the pain therapy they need. Another important new “tool” in this regard is the emergence of abuse-deterrent opioid formulations.

Cross-national comparisons of sex differences in opportunities to use alcohol or drugs, and the transitions to use

Sex differences in opportunities to use alcohol or drugs, and transition to use, were investigated in 15 surveys, in 2001-2004 (Europe 6; Americas 3; Africa 2, Asia 3; Oceania 1). The paper focuses on 18-29 year olds (N = 9,873). The World Mental Health Survey Initiative oversaw the surveys; each country obtained its own funding. A complex picture emerged with different results for alcohol and for drugs and for opportunity to use and the transition to use. Sex differences in opportunity to use alcohol were small except in Lebanon and Nigeria, whereas for drugs, the largest differences were in Mexico and Colombia.

Prediction and verification of possible reef-fish spawning aggregation sites in Los Roques Archipelago National Park, Venezuela

This study attempts to predict and verify possible spawning aggregation sites and times in the Los Roques Archipelago National Park, Venezuela, based on physical reef characteristics and the knowledge of experienced local fishermen. Three possible aggregation sites were selected for monitoring based on satellite images, low-cost bathymetric mapping and interviews with experienced local fishermen. Abundances and sizes of 18 species that are known to form reproductive aggregations were monitored at these sites using underwater visual census for 7 days after each full moon from February to August, 2007. While spawning events were not observed, possible indirect evidence of spawning aggregations was found for Lutjanus analis at Cayo Sal and Boca de Sebastopol, Lutjanus apodus at Cayo Sal, Lutjanus cyanopterus at Cayo Sal and Piedra La Guasa and Epinephelus guttatus at Bajo California and Cayo de Agua. Additionally, indirect evidence was identified for the past existence of a spawning aggregation of Epinephelus striatus in the northern part of the archipelago, which may have been eliminated by overfishing c.15 years ago. Bathymetric mapping showed that the shelf edge at sites monitored in this study was shallower than at spawning aggregation sites in other parts of the Caribbean, and that sites were not proximal to deep water. While this study does not prove the existence or locations of spawning aggregations of reef fishes in the archipelago, it does add insight to a growing understanding of generalities in the relationship between seafloor characteristics and the locations of transient reef-fish spawning aggregations in the Caribbean.

Functional analysis of the early chlorosis factor gene

Chlorosis is one of the symptoms of bacterial spot disease caused by Xanthomonas campestris pv. vesicatoria, which induces chlorosis before any other symptoms appear on tomato. We report characterization of a 2.1-kb gene called early chlorosis factor (ecf). The gene ecf encodes a hydrophobic protein with similarity to four other proteins in plant pathogens, including HolPsyAE, and uncharacterized gene products from X. campestris pv. campestris and X. axonopodis pv. citri, and, at the tertiary structure level, to colicin Ia from Escherichia coli. We demonstrate that the associated phenotype is hrp dependent, and that the ecf gene product appears to be translocated to host cells. The gene ecf has no impact on electrolyte leakage or on bacterial growth in planta in response to infection. Concentrated culture filtrates do not produce chlorosis. Study of its role in Xanthomonas spp.-tomato interactions will forward our understanding of symptom production by plant pathogens and allows further investigation into the mechanisms of bacterial virulence and production of symptoms.

Death and injury from motor vehicle crashes in Colombia

We report data on the distribution and determinants of road deaths and injuries for all victims in Colombia, with the aim of defining targets and priorities for highway death prevention in that country and other rapidly urbanizing nations. Using information from Colombia's Fund for the Prevention of Road Injury and the national death registry, we studied data on deaths and injuries from 1991 to 1995 for the nation as a whole and for the country's two largest cities, Santa Fe de Bogotá and Medellín. Deaths and injuries are rising in the nation as a whole. Of the deaths, 75% occur in urban areas, and 80% are in males. Pedestrians aged 15-34 are a peak subgroup. Thirty-four percent of deaths are attributable to speeding and/or alcohol consumption. Death tolls are highest at night and on weekends. Specific priority targets for intervention are indicated by the fact that 75% of road deaths in Colombia occur in urban areas and that 80% of all victims are males.

Transmembrane helix 7 of the endothelin B receptor regulates downstream signaling

Endothelin is a 21-amino acid peptide with a striking diversity of important biological responses, including, vasoconstriction, bronchoconstriction, and mitogenesis. Endothelin-1 binding to the endothelin B receptor (ETB), a member of the superfamily of G-protein-coupled receptors, was associated with catalytic activation of the extracellular-regulated kinase 2 (ERK2) and stimulation of AP-1 transcriptional reporter activity. A panel of single point mutations in transmembrane helix 6 (TM6), intracellular loop 3, and transmembrane helix 7 (TM7) were developed to study the structural requirements for ETB activation. Point mutations within highly conserved regions of TM6 and intracellular loop 3 were without effect on agonist-stimulated ERK activation. However, mutations within TM7 of the ETB significantly impacted ligand-stimulated downstream signaling. For example, nine point mutations within TM7 of the ETB were identified that prevented endothelin-stimulated ERK activation. Interestingly, the TM7 mutants fell into two classes; several exhibited greatly decreased AP-1 activity, relative to wild type ETB, whereas others displayed augmented endothelin-stimulated AP-1 transcriptional activity relative to wild type ETB. Our results suggest that TM7 of the ETB is involved in its activation mechanism and regulates agonist-stimulated ERK activation.

Endothelin-stimulated ERK activation in airway smooth-muscle cells requires calcium influx and Raf activation

Endothelin (ET)-1 is a 21-amino-acid peptide that is a potent vasoconstrictor and mitogen. By binding to its G-protein coupled receptor, ET-1 stimulates the proliferation of airway smooth-muscle (ASM) cells, which may be involved in the pathogenesis of asthma. The ETB receptor stimulates activation of the extracellular regulated kinase 2 (ERK2), which is thought to be required for proliferation of ASM cells. Our findings reveal that ET rapidly activates Raf, and that dominant-negative Raf interferes with ET-induced ERK activation in ASM cells. Expression of the amino-terminal Ras-binding domain of Raf inhibited ET-induced ERK activation, suggesting that ET-stimulated Raf activation is a Ras-dependent process. Furthermore, ET-stimulated ERK and Raf activation in ASM cells require calcium influx; chelating extracellular calcium or preventing calcium influx through calcium channels inhibited ET-stimulated, but not phorbol ester-stimulated, ERK and Raf activation.

Inhibition of ERK activation attenuates endothelin-stimulated airway smooth muscle cell proliferation

Endothelin is a small peptide that is a potent bronchoconstrictor, mitogen for airway smooth muscle (ASM), and is believed to be involved in the pathogenesis of asthma. To understand how endothelin stimulates the proliferation of ASM cells in culture, we evaluated the relationship between mitogen activated protein (MAP) kinase activation and cell proliferation. Endothelin is a potent stimulator of the extracellular regulated kinase 2 (ERK2) subgroup of MAP kinases, and ERK2 activation was tightly correlated with the proliferation of rat ASM cells. PD98059, a small molecule inhibitor of MEK (MAP or ERK kinase) was used to establish the role of ERK2 activation in the endothelin-stimulated signal transduction pathway leading to cell proliferation. While PD98059 significantly inhibited the ability of endothelin to activate ERK, the drug did not appear to effect the catalytic activity of an activated MEK mutant, or ERK in vitro. The data suggest that the mechanism of PD98059 inhibition of the ERK2 pathway in ASM cells may involve inhibition of MEK activation. The endothelin signal transduction pathway that culminates in ERK2 activation was dependent on protein kinase C (PKC), since depletion of PKC significantly inhibited the ability of endothelin to activate ERK2. Taken together, the data imply that activation of ERK is a critical endpoint in the endothelin signal transduction pathway since inhibition of this kinase inhibits endothelin-induced ASM cell proliferation.

Activation of multiple mitogen-activated protein kinase signal transduction pathways by the endothelin B receptor requires the cytoplasmic tail

Endothelin is a 21-amino acid peptide with remarkably diverse biological properties, including potent vasoconstriction, induction of mitogenesis, and a role in the development of blood vessels. In the present study, stimulation of the endothelin B receptor was found to activate three distinct mitogen-activated protein kinase signal transduction pathways, the extracellular regulated kinase (ERK) 2, c-Jun N-terminal kinase 1 (JNK), and p38 kinase. These mitogen-activated protein kinase isozymes are thought to mediate very different biological outcomes, suggesting that the observed pattern of kinases activation may be important for the diverse biological properties of endothelin. The cytoplasmic tail of the endothelin B receptor was found to be required for activation of all three mitogen-activated protein kinases and stimulation of intracellular calcium levels. An endothelin B receptor truncated at the C-terminal tail was not able to stimulate the mitogen-activated protein kinases or increase cytosolic free calcium. Furthermore, ectopic expression of the cytoplasmic tail attenuated signaling through the wild type receptor. The observed ERK activation appeared to be mediated by heterotrimeric G proteins, since ectopic expression of a transducin alpha-subunit inhibited endothelin-stimulated ERK activation. The data suggest that the cytosolic tail of the endothelin B receptor is involved in calcium mobilization and mitogen-activated protein kinase activation via a G protein-dependent mechanism.

Asbestos causes stimulation of the extracellular signal-regulated kinase 1 mitogen-activated protein kinase cascade after phosphorylation of the epidermal growth factor receptor

Asbestos fibers are human carcinogens with undefined mechanisms of action. In studies here, we examined signal transduction events induced by asbestos in target cells of mesothelioma and potential cell surface origins for these cascades. Asbestos fibers, but not their nonfibrous analogues, induced protracted phosphorylation of the mitogen-activated protein (MAP) kinases and extracellular signal-regulated kinases (ERK) 1 and 2, and increased kinase activity of ERK2. ERK1 and ERK2 phosphorylation and activity were initiated by addition of exogenous epidermal growth factor (EGF) and transforming growth factor-alpha, but not by isoforms of platelet-derived growth factor or insulin-like growth factor-1 in mesothelial cells. MAP kinase activation by asbestos was attenuated by suramin, which inhibits growth factor receptor interactions, or tyrphostin AG 1478, a specific inhibitor of EGF receptor tyrosine kinase activity (IC50 = 3 nM). Moreover, asbestos caused autophosphorylation of the EGF receptor, an event triggering the ERK cascade. These studies are the first to establish that a MAP kinase signal transduction pathway is initiated after phosphorylation of a peptide growth factor receptor following exposure to asbestos fibers.

p42 mitogen-activated protein kinase in brain: prominent localization in neuronal cell bodies and dendrites

Neurotransmitters and growth factors can trigger activation of a newly described family of mitogen-activated protein kinases. To help define the role of this kinase family in signal transduction in the nervous system, we have conducted immunohistochemical studies to localize p42 mitogen-activated protein kinase in rat brain sections. Light-microscopic studies revealed staining in neuronal cell bodies and dendrites that is particularly prominent in superficial layers of the neocortex, the hippocampal CA3 region and dentate gyrus, as well as cerebellar Purkinje cells. Discrete staining of oligodendrocytes was also apparent in fiber tracts, indicating expression of p42 mitogen-activated protein kinase in both neuronal and glial cell types. Electron-microscopic studies demonstrated that staining in dendrites is closely associated with microtubules. In the cell bodies, prominent staining was associated with the Golgi apparatus. In contrast, immunolabeling of synaptic terminals was not detected. Previous studies have demonstrated that p42 mitogen-activated protein kinase responds to neuronal stimulation. Immunohistochemical studies presented in this paper demonstrate prominent staining for this kinase in neuronal cell bodies and dendrites. Therefore, this kinase is likely to play a key role in postsynaptic signal transduction. As both p42 mitogen-activated protein kinase and microtubule-associated protein 2, an in vitro substrate of p42 mitogen-activated kinase, are associated with dendritic microtubules, this kinase may mediate effects of growth factors or neurotransmitters on the dendritic cytoskeleton.

Phosphorylation of mitogen-activated protein kinases is involved in the epidermal growth factor and phorbol ester, but not in the thyrotropin/cAMP, thyroid mitogenic pathway

In dog thyroid epithelial cells (thyrocytes) in primary culture, thyrotropin (TSH) acting through cAMP induces proliferation and differentiation expression, whereas epidermal growth factor (EGF) and tumor-promoting phorbol esters induce proliferation and dedifferentiation. In these cells we have demonstrated mitogen-activated protein (MAP) kinase phosphorylation by 32P labeling and two-dimensional gel electrophoresis and by immunodetection with anti-MAP kinase and anti-phosphotyrosine antibodies after one- or two-dimensional gel electrophoresis. MAP kinase localization was demonstrated by immunochemical staining. We show the following results. (i) As in other systems, EGF and phorbol esters induced p42 and p44 MAP kinases phosphorylation on tyrosine, serine, and threonine. This effect was rapid, peaking after 5 and 15 min, respectively, followed by a slow decline thereafter. It preceded a translocation of MAP kinase immunoreactivity from cytoplasm to nucleus. (ii) Carbamylcholine, a potent stimulator of the Ca(2+)-phosphatidylinositol cascade which is unable to induce DNA synthesis, stimulated MAP kinases phosphorylation and nuclear staining with kinetics similar to those observed after EGF action, indicating that MAP kinase phosphorylation was not sufficient for mitogenesis. (iii) The cAMP-dependent mitogenic cascade elicited by TSH and forskolin did not involve the phosphorylation and nuclear translocation of p42 and p44 MAP kinases at any time during the entire prereplicative phase. Activation of MAP kinases by phosphorylation is therefore not a necessary step in the G0-G1 transition in this mitogenic cascade.

Mos stimulates MAP kinase in Xenopus oocytes and activates a MAP kinase kinase in vitro

Several protein kinases, including Mos, maturation-promoting factor (MPF), mitogen-activated protein (MAP) kinase, and MAP kinase kinase (MAPKK), are activated when Xenopus oocytes enter meiosis. De novo synthesis of the Mos protein is required for progesterone-induced meiotic maturation. Recently, bacterially synthesized maltose-binding protein (MBP)-Mos fusion protein was shown to be sufficient to initiate meiosis I and MPF activation in fully grown oocytes in the absence of protein synthesis. Here we show that MAP kinase is rapidly phosphorylated and activated following injection of wild-type, but not kinase-inactive mutant, MBP-Mos into fully grown oocytes. MAP kinase activation by MBP-Mos occurs within 20 min, much more rapidly than in progesterone-treated oocytes. The MBP-Mos fusion protein also activates MPF, but MPF activation does not occur until approximately 2 h after injection. Extracts from oocytes injected with wild-type but not kinase-inactive MBP-Mos contain an activity that can phosphorylate MAP kinase, suggesting that Mos directly or indirectly activates a MAPKK. Furthermore, activated MBP-Mos fusion protein is able to phosphorylate and activate a purified, phosphatase-treated, rabbit muscle MAPKK in vitro. Thus, in oocytes, Mos is an upstream activator of MAP kinase which may function through direct phosphorylation of MAPKK.

Purification and characterization of mitogen-activated protein kinase activator(s) from epidermal growth factor-stimulated A431 cells

Two peaks of mitogen-activated protein (MAP) kinase activator activity are resolved upon ion exchange chromatography of cytosolic extracts from epidermal growth factor-stimulated A431 cells. Two forms of the activator (1 and 2) have been purified from these peaks, using chromatography on Q-Sepharose, heparin-agarose, hydroxylapatite, ATP-agarose, Sephacryl S-300, Mono S, and Mono Q. The two preparations each contained one major protein band with an apparent molecular mass of 46 or 45 kDa, respectively, on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Evidence identifying the MAP kinase activators as the 46- and 45-kDa proteins is presented. Using inactive mutants of MAP kinase as potential substrates, it was found that each preparation of MAP kinase activator catalyzes phosphorylation of the regulatory residues, threonine 188 and tyrosine 190, of Xenopus MAP kinase. These results support the concept that the MAP kinase activators are protein kinases. These MAP kinase kinases demonstrate an apparent high degree of specificity toward the native conformation of MAP kinase, although slow autophosphorylation on serine, threonine, and tyrosine residues and phosphorylation of myelin basic protein on serine and threonine residues is detected as well.

Molecular signal integration. Interplay between serine, threonine, and tyrosine phosphorylation

Images

Requirements for phosphorylation of MAP kinase during meiosis in Xenopus oocytes

Mitogen-activated protein (MAP) kinases are activated in response to a variety of extracellular stimuli by phosphorylation on tyrosine and threonine residues. Xp42 is a Xenopus laevis MAP kinase that is activated during oocyte maturation. Modified forms of Xp42 that lacked enzymatic activity or either of the phosphorylation sites were expressed in Xenopus oocytes. When meiotic maturation was induced with progesterone, each mutant Xp42 was phosphorylated, indicating that at least one kinase was activated that can phosphorylate Xp42 on tyrosine and threonine. Phosphorylation of one residue is not strictly dependent on phosphorylation of the other.

Tyrosine phosphorylation and activation of homologous protein kinases during oocyte maturation and mitogenic activation of fibroblasts

Meiotic maturation of Xenopus and sea star oocytes involves the activation of a number of protein-serine/threonine kinase activities, including a myelin basic protein (MBP) kinase. A 44-kDa MBP kinase (p44mpk) purified from mature sea star oocytes is shown here to be phosphorylated at tyrosine. Antiserum to purified sea star p44mpk was used to identify antigenically related proteins in Xenopus oocytes. Two tyrosine-phosphorylated 42-kDa proteins (p42) were detected with this antiserum in Xenopus eggs. Xenopus p42 chromatographs with MBP kinase activity on a Mono Q ion-exchange column. Tyrosine phosphorylation of Xenopus p42 approximately parallels MBP kinase activity during meiotic maturation. These results suggest that related MBP kinases are activated during meiotic maturation of Xenopus and sea star oocytes. Previous studies have suggested that Xenopus p42 is related to the mitogen-activated protein (MAP) kinases of culture mammalian cells. We have cloned a MAP kinase relative from a Xenopus ovary cDNA library and demonstrate that this clone encodes the Xenopus p42 that is tyrosine phosphorylated during oocyte maturation. Comparison of the sequences of Xenopus p42 and a rat MAP kinase (ERK1) and peptide sequences from sea star p44mpk indicates that these proteins are close relatives. The family members appear to be tyrosine phosphorylated, and activated, in different contexts, with the murine MAP kinase active during the transition from quiescence to the G1 stage of the mitotic cell cycle and the sea star and Xenopus kinases being active during M phase of the meiotic cell cycle.

Protein kinase C in adriamycin action and resistance in mouse sarcoma 180 cells

Adriamycin has a wide variety of biological actions on susceptible cells, several of which may be integrally involved in cytotoxicity. In this paper, we present evidence that one of the alterations in cell function that occurs in the presence of Adriamycin is an elevation in the production of diacylglycerol. The effect is rapid, reaches a peak within 10 min of exposure of Sarcoma 180 cells to Adriamycin, and can thus be classified among the earliest alterations that occur in cells damaged by Adriamycin. Concomitant with the rise in diacylglycerol is an increase in cytosolic protein kinase C activity. Although Adriamycin does not appear to modulate the activity of this enzyme by direct binding, drug-exposed Sarcoma 180 cells have a 56% increase in intrinsic cytosolic protein kinase C (PKC) activity, with no change in the activity of the membrane form. Experiments with the phorbol ester 12-O-tetradecanoylphorbol-13-acetate suggest that the PKC effect is linked to Adriamycin action, since activation of the enzyme by short 12-O-tetradecanoylphorbol-13-acetate exposure enhances Adriamycin's cytotoxicity as well as its ability to provoke DNA damage (measured by alkaline elution). Likewise, down-regulation of PKC by extended 12-O-tetradecanoylphorbol-13-acetate exposure partially protects the cells from Adriamycin-induced cytotoxicity as well as from DNA damage. Thus, the ability of cells to be injured by Adriamycin appears to be correlated with the activity of PKC. Multidrug-resistant subline Sarcoma 180A10 cells have the same total quantity of membrane-recruitable PKC as the sensitive parent Sarcoma 180 cells, as determined by [3H]phorbol-12,13-dibutyrate binding. However, the resistant cells have a significantly higher intrinsic PKC activity and an altered ability to translocate the enzyme to the cell surface. Taken together, the results raise the possibility that cell signaling mechanisms, particularly those involving protein kinase C, may play an important role in mediating the biological action of the anticancer drug Adriamycin.