Dr. Geoffrey Tso is a board-certified internal medicine physician who provides comprehensive care ranging from preventive health care to the management of chronic medical conditions. He strives to improve access to healthcare, and partners with his patients to provide evidence-based care tailored to the individual.

Dr. Tso is board-certified in clinical informatics and completed a postdoctoral research fellowship in Medical Informatics at the Center for Innovation to Implementation (Ci2i) at the VA in Palo Alto and the Center for Health Policy/Primary Care and Outcomes Research (CHP/PCOR) in the Stanford University School of Medicine. He has expertise in the research and implementation of generative AI technologies such as large language models. His other informatics interests include: improving patient care through clinical decision support systems, modeling knowledge for computer systems and interoperability, enhancing healthcare with AI and machine learning, using telehealth to transform how patients interface with healthcare, and innovating via digital health initiatives.

Clinical Focus

  • Internal Medicine

Academic Appointments

Professional Education

  • Board Certification, American Board of Preventative Medicine, Clinical Informatics (2018)
  • Board Certification, American Board of Internal Medicine, Internal Medicine (2015)
  • Fellowship, Stanford University School of Medicine / Palo Alto VA Healthcare System, Medical Informatics (2017)
  • Residency, UCLA-Olive View, Internal Medicine (2015)
  • Doctor of Medicine, Albert Einstein College of Medicine (2012)
  • Postbaccalaureate, University of Pennsylvania, Pre-Medicine
  • Bachelor of Science, University of California, Berkeley, Electrical Engineering and Computer Science

Current Research and Scholarly Interests

Clinical Informatics, Clinical Decision Support, Digital Health, Multimorbidity, Preventive Health, Telemedicine, Telehealth, Machine Learning, Artificial Intelligence

All Publications

  • DERA: Enhancing Large Language Model Completions with Dialog-Enabled Resolving Agents arXiv:2303.17071 Nair, V., Schumacher, E., Tso, G. J., Kannan, A. 2023
  • Interruptive Electronic Alerts for Choosing Wisely Recommendations: A Cluster Randomized Controlled Trial. Journal of the American Medical Informatics Association : JAMIA Ho, V. T., Aikens, R. C., Tso, G., Heidenreich, P. A., Sharp, C., Asch, S. M., Chen, J. H., Shah, N. K. 2022


    OBJECTIVE: To assess the efficacy of interruptive electronic alerts in improving adherence to the American Board of Internal Medicine's Choosing Wisely recommendations to reduce unnecessary laboratory testing.MATERIALS AND METHODS: We administered 5 cluster randomized controlled trials simultaneously, using electronic medical record alerts regarding prostate-specific antigen (PSA) testing, acute sinusitis treatment, vitamin D testing, carotid artery ultrasound screening, and human papillomavirus testing. For each alert, we assigned 5 outpatient clinics to an interruptive alert and 5 were observed as a control. Primary and secondary outcomes were the number of postalert orders per 100 patients at each clinic and number of triggered alerts divided by orders, respectively. Post hoc analysis evaluated whether physicians experiencing interruptive alerts reduced their alert-triggering behaviors.RESULTS: Median postalert orders per 100 patients did not differ significantly between treatment and control groups; absolute median differences ranging from 0.04 to 0.40 for PSA testing. Median alerts per 100 orders did not differ significantly between treatment and control groups; absolute median differences ranged from 0.004 to 0.03. In post hoc analysis, providers receiving alerts regarding PSA testing in men were significantly less likely to trigger additional PSA alerts than those in the control sites (Incidence Rate Ratio 0.12, 95% CI [0.03-0.52]).DISCUSSION: Interruptive point-of-care alerts did not yield detectable changes in the overall rate of undesired orders or the order-to-alert ratio between active and silent sites. Complementary behavioral or educational interventions are likely needed to improve efforts to curb medical overuse.CONCLUSION: Implementation of interruptive alerts at the time of ordering was not associated with improved adherence to 5 Choosing Wisely guidelines.TRIAL REGISTRATION: NCT02709772.

    View details for DOI 10.1093/jamia/ocac139

    View details for PubMedID 36018731

  • Open Set Medical Diagnosis Machine Learning for Healthcare (ML4H) Workshop at NeurIPS 2019 Prabhu, V., Kannan, A., Tso, G. J., Katariya, N., Chablani, M., Sontag, D., Amatriain, X. 2019
  • Learning from the experts: From expert systems to machine-learned diagnosis models Proceedings of the 3rd Machine Learning for Healthcare Conference Ravuri, M., Kannan, A., Tso, G. J., Amatriain, X. 2018; 85: 227-243
  • High-Risk Drug-Drug Interactions Between Clinical Practice Guidelines for Management of Chronic Conditions. AMIA Joint Summits on Translational Science proceedings. AMIA Joint Summits on Translational Science Tso, G. J., Tu, S. W., Musen, M. A., Goldstein, M. K. 2017; 2017: 531–39


    Clinicians and clinical decision-support systems often follow pharmacotherapy recommendations for patients based on clinical practice guidelines (CPGs). In multimorbid patients, these recommendations can potentially have clinically significant drug-drug interactions (DDIs). In this study, we describe and validate a method for programmatically detecting DDIs among CPG recommendations. The system extracts pharmacotherapy intervention recommendations from narrative CPGs, normalizes the terms, creates a mapping of drugs and drug classes, and then identifies occurrences of DDIs between CPG pairs. We used this system to analyze 75 CPGs written by authoring entities in the United States that discuss outpatient management of common chronic diseases. Using a reference list of high-risk DDIs, we identified 2198 of these DDIs in 638 CPG pairs (46 unique CPGs). Only 9 high-risk DDIs were discussed by both CPGs in a pairing. In 69 of the pairings, neither CPG had a pharmacologic reference or a warning of the possibility of a DDI.

    View details for PubMedID 28815153

  • West Nile Meningoencephalitis Presenting as Isolated Bulbar Palsy With Hypercapnic Respiratory Failure: Case Report and Literature Review. Journal of intensive care medicine Tso, G., Kaldas, K., Springer, J., Barot, N., Kamangar, N. 2016; 31 (4): 285-287


    Since the outbreak of West Nile virus (WNV) in the United States in 1999, the WNV neuroinvasive disease has been increasingly reported with a wide spectrum of neuromuscular manifestations.We submit a case of a 46-year-old male with a history of alcohol abuse, diabetes, hypertension, and hepatitis C who presented with fever, nausea, shortness of breath, and dysphagia. The patient rapidly developed hypercapnic respiratory failure and was found to have WNV meningoencephalitis without obvious neuromuscular weakness. His hospital course was significant for repeated failures of extubation secondary to persistent bulbar weakness eventually requiring tracheotomy.This is a unique case of WNV meningoencephalitis with bulbar palsy without other neuromuscular manifestations resulting in recurrent hypercapnic respiratory failure.

    View details for DOI 10.1177/0885066615589734

    View details for PubMedID 26065427

  • Test Case Selection in Pre-Deployment Testing of Complex Clinical Decision Support Systems. AMIA Joint Summits on Translational Science proceedings. AMIA Joint Summits on Translational Science Tso, G. J., Yuen, K., Martins, S., Tu, S. W., Ashcraft, M., Heidenreich, P., Hoffman, B. B., Goldstein, M. K. 2016; 2016: 240-249


    Clinical decision support (CDS) systems with complex logic are being developed. Ensuring the quality of CDS is imperative, but there is no consensus on testing standards. We tested ATHENA-HTN CDS after encoding updated hypertension guidelines into the system. A logic flow and a complexity analysis of the encoding were performed to guide testing. 100 test cases were selected to test the major pathways in the CDS logic flow, and the effectiveness of the testing was analyzed. The encoding contained 26 decision points and 3120 possible output combinations. The 100 cases selected tested all of the major pathways in the logic, but only 1% of the possible output combinations. Test case selection is one of the most challenging aspects in CDS testing and has a major impact on testing coverage. A test selection strategy should take into account the complexity of the system, identification of major logic pathways, and available resources.

    View details for PubMedID 27570678

  • Automating Guidelines for Clinical Decision Support: Knowledge Engineering and Implementation. AMIA ... Annual Symposium proceedings. AMIA Symposium Tso, G. J., Tu, S. W., Oshiro, C., Martins, S., Ashcraft, M., Yuen, K. W., Wang, D., Robinson, A., Heidenreich, P. A., Goldstein, M. K. 2016; 2016: 1189-1198


    As utilization of clinical decision support (CDS) increases, it is important to continue the development and refinement of methods to accurately translate the intention of clinical practice guidelines (CPG) into a computable form. In this study, we validate and extend the 13 steps that Shiffman et al.(5) identified for translating CPG knowledge for use in CDS. During an implementation project of ATHENA-CDS, we encoded complex CPG recommendations for five common chronic conditions for integration into an existing clinical dashboard. Major decisions made during the implementation process were recorded and categorized according to the 13 steps. During the implementation period, we categorized 119 decisions and identified 8 new categories required to complete the project. We provide details on an updated model that outlines all of the steps used to translate CPG knowledge into a CDS integrated with existing health information technology.

    View details for PubMedID 28269916

  • Normal Echocardiographic Measurements in Uncomplicated Pregnancy, a Single Center Experience Journal of Cardiovascular Disease Research Tso, G. J., et al 2014

    View details for DOI 10.5530/jcdr.2014.2.2