JBJason M. Baron

Jason M. Baron

*This is an algorithmically generated author record based on publicly available information.

Jason M. Baron has a total of 52 co-authors. Here are the top co-authors by number of shared publications:

  • Anand S. Dighe

    Department of Pathology, Massachusetts General Hospital, Harvard Medical School, 55 Fruit Street, Boston, MA 02214, USA

    4 shared publications
  • Peter Szolovits

    MIT Computer Science and Artificial Intelligence Lab , Boston, MA 02139, United States

    4 shared publications
  • Yuan Luo

    Department of Preventive Medicine Northwestern University Chicago IL USA

    3 shared publications
  • Ambarish Pandey

    Division of Cardiology, Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, Texas

    2 shared publications
  • Anand Dighe

    Department of Pathology, Massachusetts General Hospital , Boston, MA 02114, United States

    2 shared publications
  • Dagmar Schumacher

    Roche Diagnostics Corporation 3 , Indianapolis, IN

    2 shared publications

Latest Publications

  • Open Access

    Re-evaluating the Potential Impact of PrEP in Achieving HIV Elimination in the United States: Insights from Modeling

    by Jason Baron, Tamar Tchelidze, Neil Parkin, Nicole Robinson, Ben LaBrot, Aniruddha Hazra, Kenneth H Mayer

    Background: Pre-exposure prophylaxis (PrEP) to prevent HIV-1 acquisition is a critical component of the strategy to achieve epidemic control and elimination goals. Previous studies assessing PrEP needs have not considered people based on individualized, quantitative risk profiles, potentially overlooking opportunities to increase PrEP use among all individuals vulnerable to acquiring HIV. Methods: We modelled the risk distribution of sexually-acquired HIV infection based on patterns of sexual activity among sexually active adults living in the US. We estimated the population impact of PrEP use assuming “optimal” (PrEP perfectly distributed to those at highest risk) or “suboptimal” (distributed to only 50% of highest-risk individuals) PrEP distribution. Results: The model predicts that PrEP use by 1 million and 10 million individuals would prevent ∼3,400 and 13,600 sexually-acquired HIV infections per year, respectively, with optimal distribution, and 2,700 and 9,100 infections with suboptimal distribution. In addition to men who have sex with men, heterosexual women make up a significant proportion of the optimally targeted population. Conclusions: Our model provides a framework to estimate the potential population impact of PrEP. While results vary depending on assumptions, our findings strongly suggest that increases in PrEP uptake, substantially beyond current targets, may be required to achieve epidemic elimination goals. Our findings can inform updated assessments of PrEP coverage and provider initiatives at the individual level. Assessment of HIV acquisition risk based on individual behavior rather than risk-group membership indicates the imperative for wider PrEP distribution than the currently occurs.

  • Open Access

    Real-World Biomarker Test Ordering Practices in Non–Small Cell Lung Cancer: Interphysician Variation and Association With Clinical Outcomes

    by Jason M. Baron, Sarrah Widatalla, Matthew A. Gubens, Farah Khalil

    PURPOSE: Patients with metastatic or advanced non–small cell lung cancer (NSCLC) need biomarker testing, including, in most cases, anaplastic lymphoma kinase (ALK), epidermal growth factor receptor (EGFR), and PD-L1, to identify options for targeted therapies and to optimally incorporate immune checkpoint inhibitors into therapeutic regimens. We sought to examine real-world patterns of biomarker testing, quantify interphysician practice variation, and correlate testing with clinical outcomes. METHODS: We extracted real-world data from a nationwide electronic health record–derived deidentified database from 17,165 patients diagnosed with advanced NSCLC between 2018 and 2021 and receiving care in the community setting. We analyzed data using descriptive analyses, fixed- and mixed-effects logistic regression models, and proportional hazard models. RESULTS: Only 67% of all 17,165 patients and 77% of patients with nonsquamous, metastatic NSCLC had ALK, EGFR, and PD-L1 testing within 90 days of diagnosis. Later diagnosis year (2019-2021 compared with 2018) was associated with higher rates of ALK, EGFR, and PD-L1 testing; stage IIIB/C disease (compared with stage IV), squamous histology, and Black or African American race were associated with lower rates. Interphysician variation was substantial with a median odds ratio between physicians (adjusted for patient factors) of 1.78 for ALK, EGFR, and PD-L1 testing. Patients with nonsquamous, metastatic NSCLC had significantly prolonged survival if tested with all three biomarkers (median, 364 days for all three v 180 for none of the three; hazard ratio, 0.67; P < .001). CONCLUSION: Rates of biomarker testing appear suboptimal with substantial interphysician variation. Testing correlates with improved survival, although causality cannot be proven from this study. Additional work is needed to address the underlying causes of suboptimal test ordering.


  • Open Access

    High Rates of Missed HIV Testing Among Oral PrEP Users in the United States From 2018–2021: A National Assessment on Compliance With HIV Testing Recommendations of the CDC PrEP Guidelines

    by Jason Baron, Tamar Tchelidze, Benjamin LaBrot, Joseph Yao, Kenneth H Mayer, Daniel Kuritzkes, Nicole Robinson, Rupa R Patel

    Background: The US Centers for Disease Control and Prevention recommends HIV testing every 3 months in oral PrEP users. We performed a national assessment of HIV testing compliance among oral PrEP users. Methods: We analyzed 408 910 PrEP prescriptions issued to 39 809 PrEP users using a national insurance claims database that contained commercial and Medicaid claims. We identified PrEP use based on pharmacy claims and outpatient diagnostic coding. We evaluated the percentage of PrEP prescription refills without HIV testing (identified by CPT codes) within the prior 3, 6, and 12 months using time to event methods. We performed subgroup and multivariate analyses by age, gender, race, insurance type, and geography. Results: Of 39 809 persons, 36 197 were commercially insured, 3612 were Medicaid-insured, and 96% identified as male; the median age (interquartile range) was 34 (29–44) years, and the Medicaid-insured PrEP users were 24% Black/African American, 44% White, and 9% Hispanic/Latinx. Within the prior 3, 6, and 12 months, respectively, the percentage of PrEP prescription fills in individuals without HIV Ag/Ab testing was 34.3% (95% CI, 34.2%–34.5%), 23.8% (95% CI, 23.7%–23.9%), and 16.6% (95% CI, 16.4%–16.7%), and the percentage without any type of HIV test was 25.8% (95% CI, 25.6%–25.9%), 14.6% (95% CI, 14.5%–14.7%), and 7.8% (95% CI, 7.7%–7.9%). Conclusions: Approximately 1 in 3 oral PrEP prescriptions were filled in persons who had not received an HIV Ag/Ab test within the prior 3 months, with evidence of health disparities. These findings inform clinical PrEP monitoring efforts and compliance with national HIV testing guidance to monitor PrEP users.


  • Using machine learning to develop smart reflex testing protocols

    by Matthew McDermott, Anand Dighe, Peter Szolovits, Yuan Luo, Jason Baron

    Objective: Reflex testing protocols allow clinical laboratories to perform second line diagnostic tests on existing specimens based on the results of initially ordered tests. Reflex testing can support optimal clinical laboratory test ordering and diagnosis. In current clinical practice, reflex testing typically relies on simple “if-then” rules; however, this limits the opportunities for reflex testing since most test ordering decisions involve more complexity than traditional rule-based approaches would allow. Here, using the analyte ferritin as an example, we propose an alternative machine learning-based approach to “smart” reflex testing. Methods: Using deidentified patient data, we developed a machine learning model to predict whether a patient getting CBC testing will also have ferritin testing ordered. We evaluate applications of this model to reflex testing by assessing its performance in comparison to possible rule-based approaches. Results: Our underlying machine learning models performed moderately well in predicting ferritin test ordering (AUC=0.731 in reference to actual ordering) and demonstrated promising potential to underlie key clinical applications. In contrast, none of the many traditionally framed, rule-based, hypothetical reflex protocols we evaluated offered sufficient agreement with actual ordering to be clinically feasible. Using chart review, we further demonstrated that the strategic deployment of our model could avoid important ferritin test ordering errors. Conclusions: Machine learning may provide a foundation for new types of reflex testing with enhanced benefits for clinical diagnosis.


  • Open Access

    Using Machine Learning to Develop Smart Reflex Testing Protocols

    by Matthew McDermott, Anand Dighe, Peter Szolovits, Yuan Luo, Jason Baron

    Objective: Reflex testing protocols allow clinical laboratories to perform second line diagnostic tests on existing specimens based on the results of initially ordered tests. Reflex testing can support optimal clinical laboratory test ordering and diagnosis. In current clinical practice, reflex testing typically relies on simple "if-then" rules; however, this limits their scope since most test ordering decisions involve more complexity than a simple rule will allow. Here, using the analyte ferritin as an example, we propose an alternative machine learning-based approach to "smart" reflex testing with a wider scope and greater impact than traditional rule-based approaches. Methods: Using patient data, we developed a machine learning model to predict whether a patient getting CBC testing will also have ferritin testing ordered, consider applications of this model to "smart" reflex testing, and evaluate the model by comparing its performance to possible rule-based approaches. Results: Our underlying machine learning models performed moderately well in predicting ferritin test ordering and demonstrated greater suitability to reflex testing than rule-based approaches. Using chart review, we demonstrate that our model may improve ferritin test ordering. Finally, as a secondary goal, we demonstrate that ferritin test results are missing not at random (MNAR), a finding with implications for unbiased imputation of missing test results. Conclusions: Machine learning may provide a foundation for new types of reflex testing with enhanced benefits for clinical diagnosis and laboratory utilization management.