Steven Tate

All Publications

• Increased Risks of Major Cardiac Adverse Events in Stimulant Use Disorder as Compared With Other Substance Use Disorders: A Propensity-score Matching Cohort Study. Journal of addiction medicine Garel, N., Greenway, K. T., Lavin, P., Pike, C. W., Hyde, R., Jutras-Aswad, D., Tate, S. D., Lembke, A. 2025

  Abstract

  Individuals with stimulant use disorders (StSUDs) present an elevated risk of cardiovascular complications compared with the general population. However, it remains unclear whether, within the subpopulation of individuals with substance use disorders (SUDs), those specifically affected by StSUDs face even higher cardiovascular complications.We conducted a retrospective cohort study using the EVERSANA databank, spanning from January 2015 to December 2023. The EVERSANA data set comprises deidentified electronic health record data aggregated and standardized across the United States. Participants included patients diagnosed with SUDs, encompassing alcohol, cannabis, opioids, stimulants, tobacco, hallucinogens, sedative-hypnotics, or inhalants. We employed the International Classification of Disease 10th (ICD-10) version codes to define the presence of StSUD and SUD. Major adverse cardiac events (MACE) were assessed, and Cox proportional hazard ratios were adjusted using high-dimensional propensity score (hdPS) matching to account for potential confounders.Among 137,106 patients with SUD, 7706 (5.6%) had StSUD. The cohort was 50.2% female, 53.0% non-White, with a mean age of 49.1 years (SD±15). After adjustment, stimulant users exhibited significantly higher MACE rates (HR=1.37, 95% CI: 1.22-1.53, P <0.001), including an elevated risk of death (HR=1.23, 95% CI: 1.02-1.47, P =0.026).Individuals with StSUD face increased MACE compared with those with nonstimulant SUDs.

  View details for DOI 10.1097/ADM.0000000000001461

  View details for PubMedID 39961096

• Clinical entity augmented retrieval for clinical information extraction. NPJ digital medicine Lopez, I., Swaminathan, A., Vedula, K., Narayanan, S., Nateghi Haredasht, F., Ma, S. P., Liang, A. S., Tate, S., Maddali, M., Gallo, R. J., Shah, N. H., Chen, J. H. 2025; 8 (1): 45

  Abstract

  Large language models (LLMs) with retrieval-augmented generation (RAG) have improved information extraction over previous methods, yet their reliance on embeddings often leads to inefficient retrieval. We introduce CLinical Entity Augmented Retrieval (CLEAR), a RAG pipeline that retrieves information using entities. We compared CLEAR to embedding RAG and full-note approaches for extracting 18 variables using six LLMs across 20,000 clinical notes. Average F1 scores were 0.90, 0.86, and 0.79; inference times were 4.95, 17.41, and 20.08 s per note; average model queries were 1.68, 4.94, and 4.18 per note; and average input tokens were 1.1k, 3.8k, and 6.1k per note for CLEAR, embedding RAG, and full-note approaches, respectively. In conclusion, CLEAR utilizes clinical entities for information retrieval and achieves >70% reduction in token usage and inference time with improved performance compared to modern methods.

  View details for DOI 10.1038/s41746-024-01377-1

  View details for PubMedID 39828800

  View details for PubMedCentralID 4287068

• Predictability of buprenorphine-naloxone treatment retention: A multi-site analysis combining electronic health records and machine learning. Addiction (Abingdon, England) Nateghi Haredasht, F., Fouladvand, S., Tate, S., Chan, M. M., Yeow, J. J., Griffiths, K., Lopez, I., Bertz, J. W., Miner, A. S., Hernandez-Boussard, T., Chen, C. A., Deng, H., Humphreys, K., Lembke, A., Vance, L. A., Chen, J. H. 2024

  Abstract

  Opioid use disorder (OUD) and opioid dependence lead to significant morbidity and mortality, yet treatment retention, crucial for the effectiveness of medications like buprenorphine-naloxone, remains unpredictable. Our objective was to determine the predictability of 6-month retention in buprenorphine-naloxone treatment using electronic health record (EHR) data from diverse clinical settings and to identify key predictors.This retrospective observational study developed and validated machine learning-based clinical risk prediction models using EHR data.Data were sourced from Stanford University's healthcare system and Holmusk's NeuroBlu database, reflecting a wide range of healthcare settings. The study analyzed 1800 Stanford and 7957 NeuroBlu treatment encounters from 2008 to 2023 and from 2003 to 2023, respectively.Predict continuous prescription of buprenorphine-naloxone for at least 6 months, without a gap of more than 30 days. The performance of machine learning prediction models was assessed by area under receiver operating characteristic (ROC-AUC) analysis as well as precision, recall and calibration. To further validate our approach's clinical applicability, we conducted two secondary analyses: a time-to-event analysis on a single site to estimate the duration of buprenorphine-naloxone treatment continuity evaluated by the C-index and a comparative evaluation against predictions made by three human clinical experts.Attrition rates at 6 months were 58% (NeuroBlu) and 61% (Stanford). Prediction models trained and internally validated on NeuroBlu data achieved ROC-AUCs up to 75.8 (95% confidence interval [CI] = 73.6-78.0). Addiction medicine specialists' predictions show a ROC-AUC of 67.8 (95% CI = 50.4-85.2). Time-to-event analysis on Stanford data indicated a median treatment retention time of 65 days, with random survival forest model achieving an average C-index of 65.9. The top predictor of treatment retention identified included the diagnosis of opioid dependence.US patients with opioid use disorder or opioid dependence treated with buprenorphine-naloxone prescriptions appear to have a high (∼60%) treatment attrition by 6 months. Machine learning models trained on diverse electronic health record datasets appear to be able to predict treatment continuity with accuracy comparable to that of clinical experts.

  View details for DOI 10.1111/add.16587

  View details for PubMedID 38923168

• Alcohol Use Disorder Pharmacotherapy on subsequent Emergency Department Visits and Readmissions Post Hospital Discharge: A Retrospective Cohort Study Garel, N., Pike, W., Low, Y., Tate, S. WILEY. 2024: 205

  View details for Web of Science ID 001195321700040

• Trends in hallucinogen-associated emergency department visits and hospitalizations in California, USA, from 2016 to 2022. Addiction (Abingdon, England) Garel, N., Tate, S., Nash, K., Lembke, A. 2024

  Abstract

  Hallucinogens encompass a diverse range of compounds of increasing scientific and public interest. Risks associated with hallucinogen use are under-researched and poorly understood. We aimed to compare the trends in hallucinogen-associated health-care use with alcohol- and cannabis-associated health-care use.We conducted an ecological study with publicly available data on International Classification of Diseases, 10th Revision (ICD-10) diagnosis codes associated with emergency department (ED) visits and hospitalizations from the California Department of Healthcare Access and Information (HCAI). HCAI includes primary and secondary ICD-10 codes reported with ED or hospital discharge from every non-federal health-care facility licensed in California, United States, from 2016 to 2022.ICD-10 codes were classified as hallucinogen-, cannabis- or alcohol-associated if they were from the corresponding category in the ICD-10 block 'mental and behavioral disorders due to psychoactive substance use'.Observed hallucinogen-associated ED visits increased by 54% between 2016 and 2022, from 2260 visits to 3476 visits, compared with a 20% decrease in alcohol-associated ED visits and a 15% increase in cannabis-associated ED visits. The observed hallucinogen-associated hospitalizations increased by 55% during the same period, from 2556 to 3965 hospitalizations, compared with a 1% increase in alcohol-associated hospitalizations and a 1% increase in cannabis-associated hospitalizations. This rise in hallucinogenic ED visits was significantly different from the trend in cannabis-associated (P < 0.001) and alcohol-associated (P = 0.005) ED visits. The hallucinogen-associated hospitalizations trend also significantly differed when compared with cannabis- (P < 0.001) and alcohol-associated (P < 0.001) hospitalizations.Hallucinogen-associated emergency department visits and hospitalizations in California, USA, showed a large relative but small absolute increase between 2016 and 2022.

  View details for DOI 10.1111/add.16432

  View details for PubMedID 38213013

• Generative Artificial Intelligence Tools in Medicine Will Amplify Stigmatizing Language. Journal of addiction medicine Tate, S. 2023

  View details for DOI 10.1097/ADM.0000000000001237

  View details for PubMedID 37862115

• The ChatGPT therapist will see you now: Navigating generative artificial intelligence's potential in addiction medicine research and patient care. Addiction (Abingdon, England) Tate, S., Fouladvand, S., Chen, J. H., Chen, C. A. 2023

  View details for DOI 10.1111/add.16341

  View details for PubMedID 37735091

• Electronic cardiac arrest triage score best predicts mortality after intervention in patients with massive and submassive pulmonary embolism. Catheterization and cardiovascular interventions : official journal of the Society for Cardiac Angiography & Interventions Hirai, T., Tate, S., Dryer, K., Jones, D., Rosenberg, J., Nathan, S., Shah, A. P., Carey, K., Churpek, M., Edelson, D., Friant, J., Paul, J., Blair, J. E. 2018; 92 (2): 366-371

  Abstract

  To determine if the cardiac arrest triage (CART) Score would better predict poor outcomes after pharmacomechanical therapy (PMT) for massive and submassive pulmonary embolism (PE) than traditional risk scores BACKGROUND: PMT for massive and submassive PE allows for clot lysis with minimal doses of fibrinolytics. Although PMT results in improved right ventricular function, and reduced pulmonary pressures and thrombus burden, predictors of poor outcome are not well-studied.We conducted a retrospective analysis of all patients who underwent PMT for massive or submassive PE at a single institution from 2010 to 2016. The CART score and electronic CART (eCART) score, derived previously as early warning scores for hospitalized patients, were compared to pulmonary embolism severity index (PESI) comparing the area under the receiver-operator characteristic curve (AUC) for predicting 30-day mortality.We studied 61 patients (56 ±17 years, 44.0% male, 29.5% massive PE, mean PESI 114.6 ± 42.7, mean CART 13.5 ± 1.39, mean eCART 108.5 ± 28.6). Thirty-day mortality was 24.6%. Treatments included rheolytic thrombectomy (32.7%), catheter-directed thrombolysis (50.8%), ultrasound-assisted thrombolysis (32.7%), and mechanical thrombectomy (4.9%). There were no differences in outcome based on technique. The eCART and CART scores had higher AUCs compared to PESI in predicting 30-day mortality (0.84 vs 0.72 vs 0.69, P = .010). We found troponin I and pro-BNP were higher in higher eCART tertiles, however AUCs were 0.51 and 0.63, respectively for 30-day mortality when used as stand-alone predictors.Compared to PESI score, CART and eCART scores better predict mortality in massive or submassive PE patients undergoing PMT.

  View details for DOI 10.1002/ccd.27624

  View details for PubMedID 29745451

• Did the 1918 influenza cause the twentieth century cardiovascular mortality epidemic in the United States? PeerJ Tate, S., Namkung, J. J., Noymer, A. 2016; 4: e2531

  Abstract

  During most of the twentieth century, cardiovascular mortality increased in the United States while other causes of death declined. By 1958, the age-standardized death rate (ASDR) for cardiovascular causes for females was 1.84 times that for all other causes, combined (and, for males, 1.79×). Although contemporary observers believed that cardiovascular mortality would remain high, the late 1950s and early 1960s turned out to be the peak of a roughly 70-year epidemic. By 1988 for females (1986 for males), a spectacular decline had occurred, wherein the ASDR for cardiovascular causes was less than that for other causes combined. We discuss this phenomenon from a demographic point of view. We also test a hypothesis from the literature, that the 1918 influenza pandemic caused the cardiovascular mortality epidemic; we fail to find support.

  View details for DOI 10.7717/peerj.2531

  View details for PubMedID 27761328

  View details for PubMedCentralID PMC5068420