JGJack Gallifant

Jack Gallifant

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Jack Gallifant has a total of 180 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Advancing neurotech justice in youth digital mental health: insights from an interdisciplinary and cross-generational workshop

    by Craig W. McFarland, Donnella S. Comeau, Sepideh Abdi, Mahsa Alborzi Avanaki, Leo Anthony Celi, Neurotech Justice Workshop Participants, Julian Adong, Shaikha Alothman, Manal Brahimi, RuQuan Brown, et al.

    Researchers and clinicians are increasingly looking to leverage artificial intelligence (AI) and digital tools to improve psychiatric care. Of particular promise is addressing the youth mental health crisis. Yet, the introduction of AI-enabled digital technologies for psychiatric treatment of young adults raises a host of ethical, legal, and societal issues (ELSI). To provide guidance in addressing these issues, we convened a two-day meeting at the Radcliffe Institute for Advanced Study at Harvard University: Advancing Neurotech Justice in Mental Health: Insights from an Interdisciplinary and Cross-Generational Workshop . The meeting brought together a diverse cohort of 17 experts and 5 students from various fields and different countries. In partnership with the MIT Critical Data team, the workshop engaged participants in an interactive Prompt-a-Thon to explore first-hand the potential benefits, biases, and harms related to the use of Large Language Model chatbots in mental health care. This Perspective reports on five principles of digital psychiatry deployment that the workshop participants determined to be the most essential: ensuring accuracy, remaining human-centric, promoting just access, protecting privacy, and providing transparency. We place these five principles within a “Neurotech Justice” framework and discuss how guardrails can be built to promote neurotech justice in digital psychiatry.

  • Evaluating the Use of Large Language Models to Answer Patient-Facing Clinical Trial Questions

    by Jack Gallifant, Lizhou Fan, Ziqian Liao, Shan Chen, Hugo J.W.L. Aerts, Dmitriy Dligach, Ramez Kouzy, Danielle S. Bitterman

    PURPOSE: The complexity of clinical trial documentation is a significant barrier to patient understanding and informed consent. Although large language models (LLMs) show promise for simplifying this information, their tendency to generate incorrect information (hallucinate) poses serious safety risks. We aimed to systematically compare the performance of a leading proprietary LLM and a widely used open-source LLM in answering authentic patient questions about clinical trials. METHODS: We curated 349 unique patient queries from 23 authoritative oncology and regulatory Web sites. A representative subset of these questions was posed to a high-performance proprietary model (GPT-4o) and a popular open-source model (Llama-3.2-8B). Two physicians, blinded to the model's identity, independently evaluated the paired responses to assess accuracy, clarity, safety, and other quality dimensions. RESULTS: A total of 374 responses were evaluated. GPT-4o demonstrated superior reliability, with no instances of information fabrication (0 of 188 responses). In contrast, Llama-3.2-8B produced fabricated claims in 14.5% of its responses (27 of 186), including a significant error regarding the ethics of placebo controls. GPT-4o also outperformed Llama-3.2-8B across other key domains, including clarity, usefulness, and self-awareness in acknowledging uncertainty. CONCLUSION: The proprietary model, GPT-4o, was significantly more reliable and safer than the open-source Llama-3.2-8B for answering patient questions about clinical trials. These findings highlight the critical need for rigorous, comparative evaluation of LLMs before their deployment in patient-facing applications. To ensure patient safety, health care systems should pair high-performing models with structured safety guardrails and continuous monitoring.


  • Open Access

    The TRIPOD-LLM reporting guideline for studies using large language models: a Korean translation

    by Jack Gallifant, Majid Afshar, Saleem Ameen, Yindalon Aphinyanaphongs, Shan Chen, Giovanni Cacciamani, Dina Demner-Fushman, Dmitriy Dligach, Roxana Daneshjou, Chrystinne Fernandes, et al.

    The TRIPOD-LLM reporting guideline for studies using large language models: a Korean translation


  • Reliability of Large Language Model Knowledge Across Brand and Generic Cancer Drug Names

    by Jack Gallifant, Shan Chen, Sandeep K. Jain, Pedro Moreira, Umit Topaloglu, Hugo J.W.L. Aerts, Jeremy L. Warner, William G. La Cava, Danielle S. Bitterman

    PURPOSE: To evaluate the performance and consistency of large language models (LLMs) across brand and generic oncology drug names in various clinical tasks, addressing concerns about potential fluctuations in LLM performance because of subtle phrasing differences that could affect patient care. METHODS: This study evaluated three LLMs (GPT-3.5-turbo-0125, GPT-4-turbo, and GPT-4o) using drug names from HemOnc ontology. The assessment included 367 generic-to-brand and 2,516 brand-to-generic pairs, 1,000 drug-drug interaction (DDI) synthetic patient cases, and 2,438 immune-related adverse event (irAE) cases. LLMs were tested on drug name recognition, word association, DDI (DDI) detection, and irAE diagnosis using both brand and generic drug names. RESULTS: LLMs demonstrated high accuracy in matching brand and generic names (GPT-4o: 97.38% for brand, 94.71% for generic, P < .01). However, they showed significant inconsistencies in word association tasks. GPT-3.5-turbo-0125 exhibited biases favoring brand names for effectiveness (odds ratio [OR], 1.43, P < .05) and being side-effect-free (OR, 1.76, P < .05). DDI detection accuracy was poor across all models (P < .01). Consistency in irAE diagnosis varied across models. CONCLUSION: Despite high proficiency in name-matching, LLMs exhibit inconsistencies when processing brand versus generic drug names in more complex tasks. These findings highlight the need for increased awareness, improved robustness assessment methods, and the development of more consistent systems for handling nomenclature variations in clinical applications of LLMs.


  • Open Access

    Race Against the Machine Learning Courses

    by Riddhi Deshpande, Donald Mlombwa, Leo Anthony Celi, Jack Gallifant, Helen D’COuto

    Despite the rapid integration of AI in healthcare, a critical gap exists in current machine learning courses: the lack of education on identifying and mitigating bias in datasets. This oversight risks perpetuating existing health disparities through biased AI models. Analyzing 11 prominent online courses, we found only 5 addressed dataset bias, often dedicating minimal time compared to technical aspects. This paper urges course developers to prioritize education on data context, equipping learners with the tools to critically evaluate the origin, collection methods, and potential biases inherent in the data. This approach fosters the creation of fair algorithms and the incorporation of diverse data sources, ultimately mitigating the harmful effects of bias in healthcare AI. While this analysis focused on publicly available courses, it underscores the urgency of addressing bias in all healthcare machine learning education. Early intervention in algorithm development is crucial to prevent the amplification of dataset and model bias, ensuring responsible and equitable AI implementation in healthcare.

  • Open Access

    The use of large language models to enhance cancer clinical trial educational materials

    by Mingye Gao, Aman Varshney, Shan Chen, Vikram Goddla, Jack Gallifant, Patrick Doyle, Claire Novack, Maeve Dillon-Martin, Teresia Perkins, Xinrong Correia, et al.

    Background: Adequate patient awareness and understanding of cancer clinical trials is essential for trial recruitment, informed decision making, and protocol adherence. Although large language models (LLMs) have shown promise for patient education, their role in enhancing patient awareness of clinical trials remains unexplored. This study explored the performance and risks of LLMs in generating trial-specific educational content for potential participants. Methods: Generative Pretrained Transformer 4 (GPT4) was prompted to generate short clinical trial summaries and multiple-choice question-answer pairs from informed consent forms from ClinicalTrials.gov. Zero-shot learning was used for summaries, using a direct summarization, sequential extraction, and summarization approach. One-shot learning was used for question-answer pairs development. We evaluated performance through patient surveys of summary effectiveness and crowdsourced annotation of question-answer pair accuracy, using held-out cancer trial informed consent forms not used in prompt development. Results: For summaries, both prompting approaches achieved comparable results for readability and core content. Patients found summaries to be understandable and to improve clinical trial comprehension and interest in learning more about trials. The generated multiple-choice questions achieved high accuracy and agreement with crowdsourced annotators. For both summaries and multiple-choice questions, GPT4 was most likely to include inaccurate information when prompted to provide information that was not adequately described in the informed consent forms. Conclusions: LLMs such as GPT4 show promise in generating patient-friendly educational content for clinical trials with minimal trial-specific engineering. The findings serve as a proof of concept for the role of LLMs in improving patient education and engagement in clinical trials, as well as the need for ongoing human oversight.


  • Open Access

    Care Phenotypes In Critical Care

    by Ll. Weishaupt, T. Wang, J. Schamroth, P. Morandini, J. Matos, Lm Hampton, J. Gallifant, A. Fiske, N. Dundas, K. David, et al.

    The Social Determinants of Health (SDoH) have long been recognised as significant drivers of health inequalities. Within healthcare settings, large EHR datasets have increasingly enabled the use of machine learning (ML) to explore how patient background and demographic factors mediate and predict clinical outcomes. The intensive care unit (ICU) in particular provides a rich source of data for such research. However, major limitations with current approaches persist, including (i) overreliance on individual demographic labels or measures of difference, (ii) the impracticality of highly intersectional patient groups and (iii) that the underlying accuracy and validity of these demographic constructs is low. The main objective of this study was to take a novel approach, to first understand who within the ICU setting receives sub-standard care and use this to create new, objective labels based on quality of care and outcomes (‘Care Phenotypes’) when different patients interface with the health system. Using the MIMIC-IV database, we focused on highly protocolised, essential care procedures (turning, mouth care) in mechanically ventilated ICU patients. We performed a series of regression analyses to understand in which patients treatment deviated from these protocols. In a cohort of 8,919 ICU patients undergoing IMV, consistent patterns in sup-optimal protocol adherence for certain groups, notably heavier patients. Compared to equivalently sick peers, for every extra weight decile, a patient can expect a reduction of one percentile in frequency of turning care (0.0760 turning interval percentile per weight percentile, p<0.05). Furthermore, patients who receive fewer turnings should also expect to receive fewer mouth care procedures, in a quantile ratio of 1 to 5 (0.2055 mouth care interval quantile per turning interval quantile, p<0.05). The findings in this initial analysis provide support to the concept of first looking at the actual care delivered to patients to illuminate the relationship between patient demographics and outcomes of interest. The ‘Care Phenotypes’ approach has the potential to improve fairness evaluations for machine learning in healthcare, support causal inference research and play a larger role in research into healthcare disparities.

  • Top Cited Papers
    Open Access

    The TRIPOD-LLM reporting guideline for studies using large language models

    by Jack Gallifant, Majid Afshar, Saleem Ameen, Yindalon Aphinyanaphongs, Shan Chen, Giovanni Cacciamani, Dina Demner-Fushman, Dmitriy Dligach, Roxana Daneshjou, Chrystinne Fernandes, et al.

    Large language models (LLMs) are rapidly being adopted in healthcare, necessitating standardized reporting guidelines. We present transparent reporting of a multivariable model for individual prognosis or diagnosis (TRIPOD)-LLM, an extension of the TRIPOD + artificial intelligence statement, addressing the unique challenges of LLMs in biomedical applications. TRIPOD-LLM provides a comprehensive checklist of 19 main items and 50 subitems, covering key aspects from title to discussion. The guidelines introduce a modular format accommodating various LLM research designs and tasks, with 14 main items and 32 subitems applicable across all categories. Developed through an expedited Delphi process and expert consensus, TRIPOD-LLM emphasizes transparency, human oversight and task-specific performance reporting. We also introduce an interactive website (https://tripod-llm.vercel.app/) facilitating easy guideline completion and PDF generation for submission. As a living document, TRIPOD-LLM will evolve with the field, aiming to enhance the quality, reproducibility and clinical applicability of LLM research in healthcare through comprehensive reporting.


  • Open Access

    Ethical debates amidst flawed healthcare artificial intelligence metrics

    by Jack Gallifant, Danielle S. Bitterman, Leo Anthony Celi, Judy W. Gichoya, Joao Matos, Liam G. McCoy, Robin L. Pierce

    Healthcare AI faces an ethical dilemma between selective and equitable deployment, exacerbated by flawed performance metrics. These metrics inadequately capture real-world complexities and biases, leading to premature assertions of effectiveness. Improved evaluation practices, including continuous monitoring and silent evaluation periods, are crucial. To address these fundamental shortcomings, a paradigm shift in AI assessment is needed, prioritizing actual patient outcomes over conventional benchmarking.