Publications

@article{obidi2024patterns,

title = {Patterns of red blood cell utilization: Harnessing electronic health records data from the Information Standard for Blood and Transplant (ISBT) 128 system within the Biologics Effectiveness and Safety (BEST) initiative},

author = {Joyce Obidi and Gayathri Sridhar and Gra\c{c}a M Dores and Barbee Whitaker and Carlos H Villa and Emily Storch and Kinnera Chada and Lisa M Schilling and Karthik Natarajan and Paul Biondich and Andrey Soares and Matthew Spotnitz and Thomas Falconer and Saptarshi Purkayastha and Nicole L Draper and Hui-Lee Wong and Matthew Stagg and Christian Reich and Steven Anderson and Azadeh Shoaibi},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {Transfusion},

publisher = {John Wiley \& Sons, Inc. Hoboken, USA},

abstract = {Background: Current hemovigilance methods generally rely on survey data or administrative claims data utilizing billing and revenue codes, each of which has limitations. We used electronic health records (EHR) linked to blood bank data to comprehensively characterize red blood cell (RBC) utilization patterns and trends in three healthcare systems participating in the U.S. Food and Drug Administration Center for Biologics Evaluation and Research Biologics Effectiveness and Safety (BEST) initiative.



Methods: We used Information Standard for Blood and Transplant (ISBT) 128 codes linked to EHR from three healthcare systems data sources to identify and quantify RBC-transfused individuals, RBC transfusion episodes, transfused RBC units, and processing methods per year during 2012-2018.



Results: There were 577,822 RBC units transfused among 112,705 patients comprising 345,373 transfusion episodes between 2012 and 2018. Utilization in terms of RBC units and patients increased slightly in one and decreased slightly in the other two healthcare facilities. About 90% of RBC-transfused patients had 1 (~46%) or 2-5 (~42%)transfusion episodes in 2018. Among the small proportion of patients with ≥12 transfusion episodes per year, approximately 60% of episodes included only one RBC unit. All facilities used leukocyte-reduced RBCs during the study period whereas irradiated RBC utilization patterns differed across facilities.



Discussion: ISBT 128 codes and EHRs were used to observe patterns of RBC transfusion and modification methods at the unit level and patient level in three healthcare systems participating in the BEST initiative. This study shows that the ISBT 128 coding system in an EHR environment provides a feasible source for hemovigilance activities.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{abid2024ai,

title = {AI Education for Fourth-Year Medical Students: Two-Year Experience of a Web-Based, Self-Guided Curriculum and Mixed Methods Study},

author = {Areeba Abid and Avinash Murugan and Imon Banerjee and Saptarshi Purkayastha and Hari Trivedi and Judy Gichoya},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {JMIR Medical Education},

volume = {10},

pages = {e46500},

publisher = {JMIR Publications Toronto, Canada},

abstract = {Background:

Artificial intelligence (AI) and machine learning (ML) are poised to have a substantial impact in the health care space. While a plethora of web-based resources exist to teach programming skills and ML model development, there are few introductory curricula specifically tailored to medical students without a background in data science or programming. Programs that do exist are often restricted to a specific specialty.



Objective:

We hypothesized that a 1-month elective for fourth-year medical students, composed of high-quality existing web-based resources and a project-based structure, would empower students to learn about the impact of AI and ML in their chosen specialty and begin contributing to innovation in their field of interest. This study aims to evaluate the success of this elective in improving self-reported confidence scores in AI and ML. The authors also share our curriculum with other educators who may be interested in its adoption.



Methods:

This elective was offered in 2 tracks: technical (for students who were already competent programmers) and nontechnical (with no technical prerequisites, focusing on building a conceptual understanding of AI and ML). Students established a conceptual foundation of knowledge using curated web-based resources and relevant research papers, and were then tasked with completing 3 projects in their chosen specialty: a data set analysis, a literature review, and an AI project proposal. The project-based nature of the elective was designed to be self-guided and flexible to each student’s interest area and career goals. Students’ success was measured by self-reported confidence in AI and ML skills in pre and postsurveys. Qualitative feedback on students’ experiences was also collected.



Results:

This web-based, self-directed elective was offered on a pass-or-fail basis each month to fourth-year students at Emory University School of Medicine beginning in May 2021. As of June 2022, a total of 19 students had successfully completed the elective, representing a wide range of chosen specialties: diagnostic radiology (n=3), general surgery (n=1), internal medicine (n=5), neurology (n=2), obstetrics and gynecology (n=1), ophthalmology (n=1), orthopedic surgery (n=1), otolaryngology (n=2), pathology (n=2), and pediatrics (n=1). Students’ self-reported confidence scores for AI and ML rose by 66% after this 1-month elective. In qualitative surveys, students overwhelmingly reported enthusiasm and satisfaction with the course and commented that the self-direction and flexibility and the project-based design of the course were essential.



Conclusions:

Course participants were successful in diving deep into applications of AI in their widely-ranging specialties, produced substantial project deliverables, and generally reported satisfaction with their elective experience. The authors are hopeful that a brief, 1-month investment in AI and ML education during medical school will empower this next generation of physicians to pave the way for AI and ML innovation in health care.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{purkayastha2024skills,

title = {A Skills Assessment Pathways-Based Program Assessment Approach in Multidisciplinary Graduate Health Informatics},

author = {Saptarshi Purkayastha and Akshita Venkat Patri and Vedasree Yerrabandi},

url = {https://ebooks.iospress.nl/doi/10.3233/SHTI231153},

doi = {10.3233/SHTI231153},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {Studies in Health Technology and Informatics},

volume = {310},

pages = {1191\textendash1195},

abstract = {Multidisciplinary graduate education programs are hard to assess because of interdependent competencies. Students in these programs come with diverse disciplinary undergraduate degrees, and it is critical to identify knowledge gaps among these diverse learner groups to provide support to fill these gaps. Health Informatics (HI) is a multidisciplinary field in which health, technology, and social science knowledge are foundational to building HI competencies. In 2017, the American Medical Informatics Association identified ten functional domains in which HI competencies are divided. Using pre/post-semester knowledge assessment surveys of graduate students (n=60) between August 2021 to May 2022 in one of the largest graduate HI programs in the United States, we identified courses (n=9) across the curriculum that help build HI-specific competencies. Using statistical analysis, we identified three skills pathways by correlating knowledge gained with course learning objectives and used this to modify the curriculum over four semesters. These skills pathways are connected through one or two courses, where students can choose electives or, in some instances, course modules or assignments that link the skills pathways. Moreover, there is a statistically significant difference in how students gain these skills depending on their prior training, even though they take the same set of courses. Gender and other demographics did not show statistical differences in skills gained. Additionally, we found that research assistantships and internships/practicums provide additional skills not covered in our HI curriculum. Our program assessment methodology and resulting curricular changes might be relevant to HI and other multidisciplinary graduate training programs.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{kuo2024denguenet,

title = {DengueNet: Dengue Prediction using Spatiotemporal Satellite Imagery for Resource-Limited Countries},

author = {Kuan-Ting Kuo and Dana Moukheiber and Sebastian Cajas Ordonez and David Restrepo and Atika Rahman Paddo and Tsung-Yu Chen and Lama Moukheiber and Mira Moukheiber and Sulaiman Moukheiber and Saptarshi Purkayastha and others},

url = {https://openreview.net/pdf?id=eEu49z2L1A},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {arXiv preprint arXiv:2401.11114},

abstract = {Dengue fever presents a substantial challenge in developing countries where sanitation infrastructure is inadequate. The absence of comprehensive healthcare systems exacerbates the severity of dengue infections, potentially leading to life-threatening circumstances. Rapid response to dengue outbreaks is also challenging due to limited information exchange and integration. While timely dengue outbreak forecasts have the potential to prevent such outbreaks, the majority of dengue prediction studies have predominantly relied on data that impose significant burdens on individual countries for collection. In this study, our aim is to improve health equity in resource-constrained countries by exploring the effectiveness of high-resolution satellite imagery as a nontraditional and readily accessible data source. By leveraging the wealth of publicly available and easily obtainable satellite imagery, we present a scalable satellite extraction framework based on Sentinel Hub, a cloud-based computing platform. Furthermore, we introduce DengueNet, an innovative architecture that combines Vision Transformer, Radiomics, and Long Short-term Memory to extract and integrate spatiotemporal features from satellite images. This enables dengue predictions on an epi-week basis. To evaluate the effectiveness of our proposed method, we conducted experiments on five municipalities in Colombia. We utilized a dataset comprising 780 high-resolution Sentinel-2 satellite images for training and evaluation. The performance of DengueNet was assessed using the mean absolute error (MAE) metric. Across the five municipalities, DengueNet achieved an average MAE of 43.92. Our findings strongly support the efficacy of satellite imagery as a valuable resource for dengue prediction, particularly in informing public health policies within countries where manually collected data is scarce and dengue virus prevalence is severe.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{burns2024theory,

title = {Theory of radiologist interaction with instant messaging decision support tools: a sequential-explanatory study},

author = {John Lee Burns and Judy Wawira Gichoya and Marc D Kohli and Josette Jones and Saptarshi Purkayastha},

url = {https://doi.org/10.1371/journal.pdig.0000297},

doi = {10.1371/journal.pdig.0000297},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {PLOS Digital Health},

volume = {3},

number = {2},

pages = {e0000297},

publisher = {Public Library of Science San Francisco, CA USA},

abstract = {Radiology specific clinical decision support systems (CDSS) and artificial intelligence are poorly integrated into the radiologist workflow. Current research and development efforts of radiology CDSS focus on 4 main interventions, based around exam centric time points\textendashafter image acquisition, intra-report support, post-report analysis, and radiology workflow adjacent. We review the literature surrounding CDSS tools in these time points, requirements for CDSS workflow augmentation, and technologies that support clinician to computer workflow augmentation. We develop a theory of radiologist-decision tool interaction using a sequential explanatory study design. The study consists of 2 phases, the first a quantitative survey and the second a qualitative interview study. The phase 1 survey identifies differences between average users and radiologist users in software interventions using the User Acceptance of Information Technology: Toward a Unified View (UTAUT) framework. Phase 2 semi-structured interviews provide narratives on why these differences are found. To build this theory, we propose a novel solution called Radibot\textemdasha conversational agent capable of engaging clinicians with CDSS as an assistant using existing instant messaging systems supporting hospital communications. This work contributes an understanding of how radiologist-users differ from the average user and can be utilized by software developers to increase satisfaction of CDSS tools within radiology.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{venkatayogi2023seeing,

title = {From Seeing to Knowing with Artificial Intelligence: A Scoping Review of Point-of-Care Ultrasound in Low-Resource Settings},

author = {Nethra Venkatayogi and Maanas Gupta and Alaukik Gupta and Shreya Nallaparaju and Nithya Cheemalamarri and Krithika Gilari and Shireen Pathak and Krithik Vishwanath and Carel Soney and Tanisha Bhattacharya and others},

year  = {2023},

date = {2023-07-21},

urldate = {2023-01-01},

journal = {Applied Sciences},

volume = {13},

number = {14},

pages = {8427},

publisher = {MDPI},

abstract = {The utilization of ultrasound imaging for early visualization has been imperative in disease detection, especially in the first responder setting. Over the past decade, rapid advancements in the underlying technology of ultrasound have allowed for the development of portable point-of-care ultrasounds (POCUS) with handheld devices. The application of POCUS is versatile, as seen by its use in pulmonary, cardiovascular, and neonatal imaging, among many others. However, despite these advances, there is an inherent inability of translating POCUS devices to low-resource settings (LRS). To bridge these gaps, the implementation of artificial intelligence offers an interesting opportunity. Our work reviews recent applications of POCUS devices within LRS from 2016 to 2023, identifying the most commonly utilized clinical applications and areas where further innovation is needed. Furthermore, we pinpoint areas of POCUS technologies that can be improved using state-of-art artificial intelligence technologies, thus enabling the widespread adoption of POCUS devices in low-resource settings.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{owosela2023identifying,

title = {Identifying and improving the "ground truth" of race in disparities research through improved EMR data reporting. A systematic review},

author = {Babajide O Owosela and Rebecca S Steinberg and Sharon L Leslie and Leo A Celi and Saptarshi Purkayastha and Rakesh Shiradkar and Janice M Newsome and Judy W Gichoya},

doi = {10.1016/j.ijmedinf.2023.105303},

year  = {2023},

date = {2023-01-01},

urldate = {2023-01-01},

journal = {International Journal of Medical Informatics},

pages = {105303},

publisher = {Elsevier},

abstract = {Background: Studies about racial disparities in healthcare are increasing in quantity; however, they are subject to vast differences in definition, classification, and utilization of race/ethnicity data. Improved standardization of this information can strengthen conclusions drawn from studies using such data. The objective of this study is to examine how data related to race/ethnicity are recorded in research through examining articles on race/ethnicity health disparities and examine problems and solutions in data reporting that may impact overall data quality.



Methods: In this systematic review, Business Source Complete, Embase.com, IEEE Xplore, PubMed, Scopus and Web of Science Core Collection were searched for relevant articles published from 2000 to 2020. Search terms related to the concepts of electronic medical records, race/ethnicity, and data entry related to race/ethnicity were used. Exclusion criteria included articles not in the English language and those describing pediatric populations. Data were extracted from published articles. This review was organized and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement for systematic reviews.



Findings: In this systematic review, 109 full text articles were reviewed. Weaknesses and possible solutions have been discussed in current literature, with the predominant problem and solution as follows: the electronic medical record (EMR) is vulnerable to inaccuracies and incompleteness in the methods that research staff collect this data; however, improved standardization of the collection and use of race data in patient care may help alleviate these inaccuracies.



Interpretation: Conclusions drawn from large datasets concerning peoples of certain race/ethnic groups should be made cautiously, and a careful review of the methodology of each publication should be considered prior to implementation in patient care.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{kathiravelu2023telehealth,

title = {The Telehealth Dilemma\textemdashHealth-Care Deserts Meet the Internet’s Remote Regions},

author = {Pradeeban Kathiravelu and Dalibor Fonovi\'{c} and Tihana Galinac Grbac and Zachary Zaiman and Lu\'{i}s Veiga and Judy Wawira Gichoya and Saptarshi Purkayastha and Babak Mahmoudi},

year  = {2023},

date = {2023-01-01},

journal = {Computer},

volume = {56},

number = {9},

pages = {39\textendash49},

publisher = {IEEE},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{burns2023ability,

title = {Ability of artificial intelligence to identify self-reported race in chest x-ray using pixel intensity counts},

author = {John Lee Burns and Zachary Zaiman and Jack Vanschaik and Gaoxiang Luo and Le Peng and Brandon Price and Garric Mathias and Vijay Mittal and Akshay Sagane and Christopher Tignanelli and Sunandan Chakraborty and Judy Wawira Gichoya and Saptarshi Purkayastha},

doi = {10.1117/1.JMI.10.6.061106},

year  = {2023},

date = {2023-01-01},

urldate = {2023-01-01},

journal = {Journal of Medical Imaging},

volume = {10},

number = {6},

pages = {061106\textendash061106},

publisher = {Society of Photo-Optical Instrumentation Engineers},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{banerjee2023shortcuts,

title = {“Shortcuts” causing bias in radiology artificial intelligence: causes, evaluation and mitigation.},

author = {Imon Banerjee and Kamanasish Bhattacharjee and John L Burns and Hari Trivedi and Saptarshi Purkayastha and Laleh Seyyed-Kalantari and Bhavik N Patel and Rakesh Shiradkar and Judy Gichoya},

doi = {10.1016/j.jacr.2023.06.025},

year  = {2023},

date = {2023-01-01},

urldate = {2023-01-01},

journal = {Journal of the American College of Radiology},

publisher = {Elsevier},

abstract = {Despite the expert-level performance of artificial intelligence (AI) models for various medical imaging tasks, real-world performance failures with disparate outputs for various subgroups limit the usefulness of AI in improving patients’ lives. Many definitions of fairness have been proposed, with discussions of various tensions that arise in the choice of an appropriate metric to use to evaluate bias; for example, should one aim for individual or group fairness? One central observation is that AI models apply “shortcut learning” whereby spurious features (such as chest tubes and portable radiographic markers on intensive care unit chest radiography) on medical images are used for prediction instead of identifying true pathology. Moreover, AI has been shown to have a remarkable ability to detect protected attributes of age, sex, and race, while the same models demonstrate bias against historically underserved subgroups of age, sex, and race in disease diagnosis. Therefore, an AI model may take shortcut predictions from these correlations and subsequently generate an outcome that is biased toward certain subgroups even when protected attributes are not explicitly used as inputs into the model. As a result, these subgroups became nonprivileged subgroups. In this review, the authors discuss the various types of bias from shortcut learning that may occur at different phases of AI model development, including data bias, modeling bias, and inference bias. The authors thereafter summarize various tool kits that can be used to evaluate and mitigate bias and note that these have largely been applied to nonmedical domains and require more evaluation for medical AI. The authors then summarize current techniques for mitigating bias from preprocessing (data-centric solutions) and during model development (computational solutions) and postprocessing (recalibration of learning). Ongoing legal changes where the use of a biased model will be penalized highlight the necessity of understanding, detecting, and mitigating biases from shortcut learning and will require diverse research teams looking at the whole AI pipeline.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{guo2023medshift,

title = {MedShift: Automated Identification of Shift Data for Medical Image Dataset Curation},

author = {Xiaoyuan Guo and Judy Wawira Gichoya and Hari Trivedi and Saptarshi Purkayastha and Imon Banerjee},

year  = {2023},

date = {2023-01-01},

urldate = {2023-01-01},

journal = {IEEE Journal of Biomedical and Health Informatics},

publisher = {IEEE},

abstract = {Automated curation of noisy external data in the medical domain has long been in high demand, as AI technologies need to be validated using various sources with clean, annotated data. Identifying the variance between internal and external sources is a fundamental step in curating a high-quality dataset, as the data distributions from different sources can vary significantly and subsequently affect the performance of AI models. The primary challenges for detecting data shifts are - (1) accessing private data across healthcare institutions for manual detection and (2) the lack of automated approaches to learn efficient shift-data representation without training samples. To overcome these problems, we propose an automated pipeline called MedShift to detect top-level shift samples and evaluate the significance of shift data without sharing data between internal and external organizations. MedShift employs unsupervised anomaly detectors to learn the internal distribution and identify samples showing significant shiftness for external datasets, and then compares their performance. To quantify the effects of detected shift data, we train a multi-class classifier that learns internal domain knowledge and evaluates the classification performance for each class in external domains after dropping the shift data. We also propose a data quality metric to quantify the dissimilarity between internal and external datasets. We verify the efficacy of MedShift using musculoskeletal radiographs (MURA) and chest X-ray datasets from multiple external sources. Our experiments show that our proposed shift data detection pipeline can be beneficial for medical centers to curate high-quality datasets more efficiently.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{li2023ethics,

title = {Ethics of large language models in medicine and medical research},

author = {Hanzhou Li and John T Moon and Saptarshi Purkayastha and Leo Anthony Celi and Hari Trivedi and Judy W Gichoya},

year  = {2023},

date = {2023-01-01},

journal = {The Lancet Digital Health},

volume = {5},

number = {6},

pages = {e333\textendashe335},

publisher = {Elsevier},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{tummala2022development,

title = {Development and evaluation of a natural language conversational bot for identifying appropriate clinician referral from patient narratives},

author = {Sriharsha Tummala and Saptarshi Purkayastha and Josette Jones},

year  = {2022},

date = {2022-01-01},

urldate = {2022-01-01},

abstract = {Recent years have seen a significant increase in automated conversational agent chatbots. Conversational agents like chatbots for health may provide timely and cost-effective support in clinical care. Some studies show that chatbots could have an impact on patient engagement. Additionally, health systems are attempting to connect with patients over social networks, mainly where specialists are limited. By 2025, the Association of American Medical Colleges estimates that the United States will have a shortfall of 61,700-94,700 physicians and critical shortage in many specialties, delaying available appointments by months in many cases. Thus, we need innovative solutions that can manage the time of limited specialists appropriately. Recent research has demonstrated that deep learning methods are superior for natural language classification tasks compared to other machine learning methods. The primary objective of this study was to develop a telegram chatbot which reads patient narratives and acts as a conversational agent by redirecting the case to the appropriate specialist. Besides simply working on improving conversational capabilities of chatbots, we developed a novel method for referring the cases to specialists based on their responses to previous cases on a social network group. As far as we know, no other chatbot has the level of accuracy or referral system like our developed chatbot.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{gichoya2022ai,

title = {AI recognition of patient race in medical imaging: a modelling study},

author = {Judy Wawira Gichoya and Imon Banerjee and Ananth Reddy Bhimireddy and John L Burns and Leo Anthony Celi and Li-Ching Chen and Ramon Correa and Natalie Dullerud and Marzyeh Ghassemi and Shih-Cheng Huang and others},

year  = {2022},

date = {2022-01-01},

urldate = {2022-01-01},

journal = {The Lancet Digital Health},

volume = {4},

number = {6},

pages = {e406\textendashe414},

publisher = {Elsevier},

abstract = {Background

Previous studies in medical imaging have shown disparate abilities of artificial intelligence (AI) to detect a person\'s race, yet there is no known correlation for race on medical imaging that would be obvious to human experts when interpreting the images. We aimed to conduct a comprehensive evaluation of the ability of AI to recognise a patient\'s racial identity from medical images.



Methods

Using private (Emory CXR, Emory Chest CT, Emory Cervical Spine, and Emory Mammogram) and public (MIMIC-CXR, CheXpert, National Lung Cancer Screening Trial, RSNA Pulmonary Embolism CT, and Digital Hand Atlas) datasets, we evaluated, first, performance quantification of deep learning models in detecting race from medical images, including the ability of these models to generalise to external environments and across multiple imaging modalities. Second, we assessed possible confounding of anatomic and phenotypic population features by assessing the ability of these hypothesised confounders to detect race in isolation using regression models, and by re-evaluating the deep learning models by testing them on datasets stratified by these hypothesised confounding variables. Last, by exploring the effect of image corruptions on model performance, we investigated the underlying mechanism by which AI models can recognise race.



Findings

In our study, we show that standard AI deep learning models can be trained to predict race from medical images with high performance across multiple imaging modalities, which was sustained under external validation conditions (x-ray imaging [area under the receiver operating characteristics curve (AUC) range 0·91\textendash0·99], CT chest imaging [0·87\textendash0·96], and mammography [0·81]). We also showed that this detection is not due to proxies or imaging-related surrogate covariates for race (eg, performance of possible confounders: body-mass index [AUC 0·55], disease distribution [0·61], and breast density [0·61]). Finally, we provide evidence to show that the ability of AI deep learning models persisted over all anatomical regions and frequency spectrums of the images, suggesting the efforts to control this behaviour when it is undesirable will be challenging and demand further study.



Interpretation

The results from our study emphasise that the ability of AI deep learning models to predict self-reported race is itself not the issue of importance. However, our finding that AI can accurately predict self-reported race, even from corrupted, cropped, and noised medical images, often when clinical experts cannot, creates an enormous risk for all model deployments in medical imaging.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{ochoa2022ntd,

title = {NTD Health: An electronic medical record system for neglected tropical diseases},

author = {Rodrigo Ochoa and Alessa \'{A}lvarez and Jordan Freitas and Saptarshi Purkayastha and Iv\'{a}n D V\'{e}lez},

year  = {2022},

date = {2022-01-01},

urldate = {2022-01-01},

journal = {Biom\'{e}dica},

volume = {42},

number = {4},

pages = {602\textendash610},

publisher = {Instituto Nacional de Salud},

abstract = {Introduction:

The use of technological resources to support processes in health systems has generated robust, interoperable, and dynamic platforms. In the case of institutions working with neglected tropical diseases, there is a need for specific customizations of these diseases.



Objectives:

To establish a medical record platform specialized in neglected tropical diseases which could facilitate the analysis of treatment evolution in patients, as well as generate more accurate data about various clinical aspects.



Materials and methods:

A set of requirements to develop state of the art forms, concepts, and functionalities to include neglected tropical diseases were compiled. An OpenMRS distribution (version 2.3) was used as reference to build the platform, following the recommended guidelines and shared-community modules.



Results:

All the customized information was developed in a platform called NTD Health, which is web-based and can be upgraded and improved by users without technological barriers.



Conclusions:

The electronic medical record system can become a useful tool for other institutions to improve their health practices as well as the quality of life for neglected tropical disease patients, simplifying the customization of healthcare systems able to interoperate with other platforms.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{ezenwa2022using,

title = {Using mobile virtual reality simulation to prepare for in-person helping babies breathe training: Secondary analysis of a randomized controlled trial (the eHBB/mHBS trial)},

author = {Beatrice Nkolika Ezenwa and Rachel Umoren and Iretiola Bamikeolu Fajolu and Daniel S Hippe and Sherri Bucher and Saptarshi Purkayastha and Felicitas Okwako and Fabian Esamai and John B Feltner and Olubukola Olawuyi and others},

year  = {2022},

date = {2022-01-01},

urldate = {2022-01-01},

journal = {JMIR Medical Education},

volume = {8},

number = {3},

pages = {e37297},

publisher = {JMIR Publications Inc., Toronto, Canada},

abstract = {Background: Neonatal mortality accounts for approximately 46% of global under-5 child mortality. The widespread access to mobile devices in low- and middle-income countries has enabled innovations, such as mobile virtual reality (VR), to be leveraged in simulation education for health care workers.



Objective: This study explores the feasibility and educational efficacy of using mobile VR for the precourse preparation of health care professionals in neonatal resuscitation training.



Methods: Health care professionals in obstetrics and newborn care units at 20 secondary and tertiary health care facilities in Lagos, Nigeria, and Busia, Western Kenya, who had not received training in Helping Babies Breathe (HBB) within the past 1 year were randomized to access the electronic HBB VR simulation and digitized HBB Provider\'s Guide (VR group) or the digitized HBB Provider\'s Guide only (control group). A sample size of 91 participants per group was calculated based on the main study protocol that was previously published. Participants were directed to use the electronic HBB VR simulation and digitized HBB Provider\'s Guide or the digitized HBB Provider\'s Guide alone for a minimum of 20 minutes. HBB knowledge and skills assessments were then conducted, which were immediately followed by a standard, in-person HBB training course that was led by study staff and used standard HBB evaluation tools and the Neonatalie Live manikin (Laerdal Medical).



Results: A total of 179 nurses and midwives participated (VR group: n=91; control group: n=88). The overall performance scores on the knowledge check (P=.29), bag and mask ventilation skills check (P=.34), and Objective Structured Clinical Examination A checklist (P=.43) were similar between groups, with low overall pass rates (6/178, 3.4% of participants). During the Objective Structured Clinical Examination A test, participants in the VR group performed better on the critical step of positioning the head and clearing the airway (VR group: 77/90, 86%; control group: 57/88, 65%; P=.002). The median percentage of ventilations that were performed via head tilt, as recorded by the Neonatalie Live manikin, was also numerically higher in the VR group (75%, IQR 9%-98%) than in the control group (62%, IQR 13%-97%), though not statistically significantly different (P=.35). Participants in the control group performed better on the identifying a helper and reviewing the emergency plan step (VR group: 7/90, 8%; control group: 16/88, 18%; P=.045) and the washing hands step (VR group: 20/90, 22%; control group: 32/88, 36%; P=.048).



Conclusions: The use of digital interventions, such as mobile VR simulations, may be a viable approach to precourse preparation in neonatal resuscitation training for health care professionals in low- and middle-income countries.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{guo2022multi,

title = {Multi-label Medical Image Retrieval via Learning Multi-class Similarity},

author = {Xiaoyuan Guo and Jiali Duan and Judy Gichoya and Hari Trivedi and Saptarshi Purkayastha and Ashish Sharma and Imon Banerjee},

year  = {2022},

date = {2022-01-01},

urldate = {2022-01-01},

publisher = {AIIM-D-22-00928},

abstract = {Introduction: Multi-label image retrieval is a challenging problem in the medical area. First, compared to natural images, labels in the medical domain exhibit higher class-imbalance and much nuanced variations. Second, pair-based sampling for positives and negatives during similarity optimization are ambiguous in the multi-label setting, as samples with the same set of labels are limited. 

Methods: To address the aforementioned challenges, we propose a proxy-based multi-class similarity (PMS) framework, which compares and contrasts samples by comparing their similarities with the discovered proxies. In this way, samples of different sets of label attributes can be utilized and compared indirectly, without the need for complicated sampling. PMS learns a class-wise feature decomposition and maintains a memory bank for positive features from each class. The memory bank keeps track of the latest features, used to compute the class proxies. We compare samples based on their similarity distributions against the proxies, which provide a more stable mean against noise.

Results: We benchmark over 10 popular metric learning baselines on two public chest X-ray datasets and experiments show consistent stability of our approach under both exact and non-exact match settings.

Conclusions: We proposed a methodology for multi-label medical image retrieval and design a proxy-based multi-class similarity metric, which compares and contrasts samples based on their similarity distributions with respect to the class proxies. With no perquisites, the metrics can be applied to various multi-label medical image applications. The implementation code repository will be publicly available after acceptance.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{guo2021cvad,

title = {CVAD/An unsupervised image anomaly detector},

author = {Xiaoyuan Guo and Judy Wawira Gichoya and Saptarshi Purkayastha and Imon Banerjee},

doi = {10.1016/j.simpa.2021.100195},

year  = {2021},

date = {2021-12-22},

urldate = {2021-01-01},

journal = {Software Impacts},

pages = {100195},

publisher = {Elsevier},

abstract = {Detecting out-of-distribution samples for image applications plays an important role in safeguarding the reliability of machine learning model deployment. In this article, we developed a software tool to support our OOD detector CVAD - a self-supervised Cascade Variational autoencoder-based Anomaly Detector , which can be easily applied to various image applications without any assumptions. The corresponding open-source software is published for better public research and tool usage.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{mahajan2021plhi,

title = {PLHI-MC10: A dataset of exercise activities captured through a triple synchronous medically-approved sensor},

author = {Yohan Mahajan and Ananth Bhimireddy and Areeba Abid and Judy W Gichoya and Saptarshi Purkayastha},

doi = {10.1016/j.dib.2021.107287},

year  = {2021},

date = {2021-10-01},

urldate = {2021-01-01},

journal = {Data in Brief},

volume = {38},

pages = {107287},

publisher = {Elsevier},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{umoren2021ehbbb,

title = {eHBB: a randomised controlled trial of virtual reality or video for neonatal resuscitation refresher training in healthcare workers in resource-scarce settings},

author = {Rachel Umoren and Sherri Bucher and Daniel S Hippe and Beatrice Nkolika Ezenwa and Iretiola Bamikeolu Fajolu and Felicitas M Okwako and John Feltner and Mary Nafula and Annet Musale and Olubukola A Olawuyi and others},

doi = {10.1136/bmjopen-2020-048506},

year  = {2021},

date = {2021-08-25},

urldate = {2021-01-01},

journal = {BMJ open},

volume = {11},

number = {8},

pages = {e048506},

publisher = {British Medical Journal Publishing Group},

abstract = {Objective To assess the impact of mobile virtual reality (VR) simulations using electronic Helping Babies Breathe (eHBB) or video for the maintenance of neonatal resuscitation skills in healthcare workers in resource-scarce settings.



Design Randomised controlled trial with 6-month follow-up (2018\textendash2020).



Setting Secondary and tertiary healthcare facilities.



Participants 274 nurses and midwives assigned to labour and delivery, operating room and newborn care units were recruited from 20 healthcare facilities in Nigeria and Kenya and randomised to one of three groups: VR (eHBB+digital guide), video (video+digital guide) or control (digital guide only) groups before an in-person HBB course.



Intervention(s) eHBB VR simulation or neonatal resuscitation video.



Main outcome(s) Healthcare worker neonatal resuscitation skills using standardised checklists in a simulated setting at 1 month, 3 months and 6 months.



Results Neonatal resuscitation skills pass rates were similar among the groups at 6-month follow-up for bag-and-mask ventilation (BMV) skills check (VR 28%, video 25%, control 22%, p=0.71), objective structured clinical examination (OSCE) A (VR 76%, video 76%, control 72%, p=0.78) and OSCE B (VR 62%, video 60%, control 49%, p=0.18). Relative to the immediate postcourse assessments, there was greater retention of BMV skills at 6 months in the VR group (−15% VR, p=0.10; −21% video, p\<0.01, \textendash27% control, p=0.001). OSCE B pass rates in the VR group were numerically higher at 3 months (+4%, p=0.64) and 6 months (+3%, p=0.74) and lower in the video (−21% at 3 months, p\<0.001; −14% at 6 months, p=0.066) and control groups (−7% at 3 months, p=0.43; −14% at 6 months, p=0.10). On follow-up survey, 95% (n=65) of respondents in the VR group and 98% (n=82) in the video group would use their assigned intervention again.



Conclusion eHBB VR training was highly acceptable to healthcare workers in low-income to middle-income countries and may provide additional support for neonatal resuscitation skills retention compared with other digital interventions.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{kathiravelu2021dicom,

title = {A DICOM Framework for Machine Learning and Processing Pipelines Against Real-time Radiology Images},

author = {Pradeeban Kathiravelu and Puneet Sharma and Ashish Sharma and Imon Banerjee and Hari Trivedi and Saptarshi Purkayastha and Priyanshu Sinha and Alexandre Cadrin-Chenevert and Nabile Safdar and Judy Wawira Gichoya},

doi = {10.1007/s10278-021-00491-w},

year  = {2021},

date = {2021-08-01},

urldate = {2021-01-01},

journal = {Journal of Digital Imaging},

volume = {34},

number = {4},

pages = {1005--1013},

publisher = {Springer International Publishing},

abstract = {Real-time execution of machine learning (ML) pipelines on radiology images is difficult due to limited computing resources in clinical environments, whereas running them in research clusters requires efficient data transfer capabilities. We developed Niffler, an open-source Digital Imaging and Communications in Medicine (DICOM) framework that enables ML and processing pipelines in research clusters by efficiently retrieving images from the hospitals’ PACS and extracting the metadata from the images. We deployed Niffler at our institution (Emory Healthcare, the largest healthcare network in the state of Georgia) and retrieved data from 715 scanners spanning 12 sites, up to 350 GB/day continuously in real-time as a DICOM data stream over the past 2 years. We also used Niffler to retrieve images bulk on-demand based on user-provided filters to facilitate several research projects. This paper presents the architecture and three such use cases of Niffler. First, we executed an IVC filter detection and segmentation pipeline on abdominal radiographs in real-time, which was able to classify 989 test images with an accuracy of 96.0%. Second, we applied the Niffler Metadata Extractor to understand the operational efficiency of individual MRI systems based on calculated metrics. We benchmarked the accuracy of the calculated exam time windows by comparing Niffler against the Clinical Data Warehouse (CDW). Niffler accurately identified the scanners’ examination timeframes and idling times, whereas CDW falsely depicted several exam overlaps due to human errors. Third, with metadata extracted from the images by Niffler, we identified scanners with misconfigured time and reconfigured five scanners. Our evaluations highlight how Niffler enables real-time ML and processing pipelines in a research cluster.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{guo2021margin,

title = {Margin-Aware Intra-Class Novelty Identification for Medical Images},

author = {Xiaoyuan Guo and Judy Wawira Gichoya and Saptarshi Purkayastha and Imon Banerjee},

url = {https://arxiv.org/abs/2108.00117},

year  = {2021},

date = {2021-07-31},

urldate = {2021-07-31},

journal = {arXiv preprint arXiv:2108.00117},

abstract = {Traditional anomaly detection methods focus on detecting inter-class variations while medical image novelty identification is inherently an intra-class detection problem. For example, a machine learning model trained with normal chest X-ray and common lung abnormalities, is expected to discover and flag idiopathic pulmonary fibrosis which a rare lung disease and unseen by the model during training. The nuances from intra-class variations and lack of relevant training data in medical image analysis pose great challenges for existing anomaly detection methods. To tackle the challenges, we propose a hybrid model - Transformation-based Embedding learning for Novelty Detection (TEND) which without any out-of-distribution training data, performs novelty identification by combining both autoencoder-based and classifier-based method. With a pre-trained autoencoder as image feature extractor, TEND learns to discriminate the feature embeddings of in-distribution data from the transformed counterparts as fake out-of-distribution inputs. To enhance the separation, a distance objective is optimized to enforce a margin between the two classes. Extensive experimental results on both natural image datasets and medical image datasets are presented and our method out-performs state-of-the-art approaches. },

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{banerjee2021reading,

title = {Reading Race: AI Recognises Patient's Racial Identity In Medical Images},

author = {Imon Banerjee and Ananth Reddy Bhimireddy and John L Burns and Leo Anthony Celi and Li-Ching Chen and Ramon Correa and Natalie Dullerud and Marzyeh Ghassemi and Shih-Cheng Huang and Po-Chih Kuo and others},

url = {https://arxiv.org/abs/2107.10356},

year  = {2021},

date = {2021-07-21},

urldate = {2021-07-21},

journal = {arXiv preprint arXiv:2107.10356},

abstract = {Background: In medical imaging, prior studies have demonstrated disparate AI performance by race, yet there is no known correlation for race on medical imaging that would be obvious to the human expert interpreting the images.

Methods: Using private and public datasets we evaluate: A) performance quantification of deep learning models to detect race from medical images, including the ability of these models to generalize to external environments and across multiple imaging modalities, B) assessment of possible confounding anatomic and phenotype population features, such as disease distribution and body habitus as predictors of race, and C) investigation into the underlying mechanism by which AI models can recognize race.

Findings: Standard deep learning models can be trained to predict race from medical images with high performance across multiple imaging modalities. Our findings hold under external validation conditions, as well as when models are optimized to perform clinically motivated tasks. We demonstrate this detection is not due to trivial proxies or imaging-related surrogate covariates for race, such as underlying disease distribution. Finally, we show that performance persists over all anatomical regions and frequency spectrum of the images suggesting that mitigation efforts will be challenging and demand further study.

Interpretation: We emphasize that model ability to predict self-reported race is itself not the issue of importance. However, our findings that AI can trivially predict self-reported race -- even from corrupted, cropped, and noised medical images -- in a setting where clinical experts cannot, creates an enormous risk for all model deployments in medical imaging: if an AI model secretly used its knowledge of self-reported race to misclassify all Black patients, radiologists would not be able to tell using the same data the model has access to. },

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{tariq2021patient,

title = {Patient-specific COVID-19 resource utilization prediction using fusion AI model},

author = {Amara Tariq and Leo Anthony Celi and Janice M Newsome and Saptarshi Purkayastha and Neal Kumar Bhatia and Hari Trivedi and Judy Wawira Gichoya and Imon Banerjee},

doi = {10.1038/s41746-021-00461-0},

year  = {2021},

date = {2021-06-03},

urldate = {2021-01-01},

journal = {NPJ digital medicine},

volume = {4},

number = {1},

pages = {1--9},

publisher = {Nature Publishing Group},

abstract = {The strain on healthcare resources brought forth by the recent COVID-19 pandemic has highlighted the need for efficient resource planning and allocation through the prediction of future consumption. Machine learning can predict resource utilization such as the need for hospitalization based on past medical data stored in electronic medical records (EMR). We conducted this study on 3194 patients (46% male with mean age 56.7 (±16.8), 56% African American, 7% Hispanic) flagged as COVID-19 positive cases in 12 centers under Emory Healthcare network from February 2020 to September 2020, to assess whether a COVID-19 positive patient’s need for hospitalization can be predicted at the time of RT-PCR test using the EMR data prior to the test. Five main modalities of EMR, i.e., demographics, medication, past medical procedures, comorbidities, and laboratory results, were used as features for predictive modeling, both individually and fused together using late, middle, and early fusion. Models were evaluated in terms of precision, recall, F1-score (within 95% confidence interval). The early fusion model is the most effective predictor with 84% overall F1-score [CI 82.1\textendash86.1]. The predictive performance of the model drops by 6 % when using recent clinical data while omitting the long-term medical history. Feature importance analysis indicates that history of cardiovascular disease, emergency room visits in the past year prior to testing, and demographic factors are predictive of the disease trajectory. We conclude that fusion modeling using medical history and current treatment data can forecast the need for hospitalization for patients infected with COVID-19 at the time of the RT-PCR test.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{purkayastha2021failuresb,

title = {Failures hiding in success for artificial intelligence in radiology},

author = {Saptarshi Purkayastha and Hari Trivedi and Judy Wawira Gichoya},

doi = {10.1016/j.jacr.2020.11.008},

year  = {2021},

date = {2021-03-01},

urldate = {2021-01-01},

journal = {Journal of the American College of Radiology},

volume = {18},

number = {3},

pages = {517--519},

publisher = {Elsevier},

abstract = {Reports of computer algorithms outperforming radiologists have persisted over the last 15 years, starting with the 2005 publication by Rubin et al on detecting pulmonary nodules from CT scans [ 1

]. Back then, these technologies were referred to as computer-aided diagnosis, which could be considered as a precursor, of sorts, to what is now referred to broadly as artificial intelligence (AI). Technology gains in hardware over the past 5 years have facilitated the training of deep neural networks with millions of parameters, exponentially accelerating the pace of AI publications. However, like every other scientific field, successes of AI in radiology are published and publicized with much fanfare, and failures are not discussed or made public. In fact, most AI failures are discovered anecdotally from personal experience or when shared in social media as tweets or blog posts. In this article, we discuss some pitfalls frequently encountered in reporting the success of AI in radiology, which might be considered failures when considered differently.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{purkayastha2021failures,

title = {Failures hiding in success for artificial intelligence in radiology},

author = {Saptarshi Purkayastha and Hari Trivedi and Judy Wawira Gichoya},

doi = {10.1016/j.jacr.2020.11.008},

year  = {2021},

date = {2021-03-01},

journal = {Journal of the American College of Radiology},

volume = {18},

number = {3},

pages = {517-519},

abstract = {Reports of computer algorithms outperforming radiologists have persisted over the last 15 years, starting with the 2005 publication by Rubin et al on detecting pulmonary nodules from CT scans. Back then, these technologies were referred to as computer-aided diagnosis, which could be considered as a precursor, of sorts, to what is now referred to broadly as artificial intelligence (AI). Technology gains in hardware over the past 5 years have facilitated the training of deep neural networks with millions of parameters, exponentially accelerating the pace of AI publications. However, like every other scientific field, the successes of AI in radiology are published and publicized with much fanfare, and failures are not discussed or made public. In fact, most AI failures are discovered anecdotally from personal experience or when shared in social media as tweets or blog posts. In this article, we discuss some pitfalls frequently encountered in reporting the success of AI in radiology, which might be considered failures when considered differently. },

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{tariq2020current,

title = {Current clinical applications of artificial intelligence in radiology and their best supporting evidence},

author = {Amara Tariq and Saptarshi Purkayastha and Geetha Priya Padmanaban and Elizabeth Krupinski and Hari Trivedi and Imon Banerjee and Judy Wawira Gichoya},

doi = {10.1016/j.jacr.2020.08.018},

year  = {2020},

date = {2020-11-02},

urldate = {2020-11-02},

journal = {Journal of the American College of Radiology},

volume = {17},

number = {11},

pages = {1371-1381},

publisher = {Elsevier},

abstract = {Purpose

Despite tremendous gains from deep learning and the promise of artificial intelligence (AI) in medicine to improve diagnosis and save costs, there exists a large translational gap to implement and use AI products in real-world clinical situations. Adoption of standards such as Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis, Consolidated Standards of Reporting Trials, and the Checklist for Artificial Intelligence in Medical Imaging is increasing to improve the peer-review process and reporting of AI tools. However, no such standards exist for product-level review.



Methods

A review of clinical trials showed a paucity of evidence for radiology AI products; thus, the authors developed a 10-question assessment tool for reviewing AI products with an emphasis on their validation and result dissemination. The assessment tool was applied to commercial and open-source algorithms used for diagnosis to extract evidence on the clinical utility of the tools.



Results

There is limited technical information on methodologies for FDA-approved algorithms compared with open-source products, likely because of intellectual property concerns. Furthermore, FDA-approved products use much smaller data sets compared with open-source AI tools, because the terms of use of public data sets are limited to academic and noncommercial entities, which precludes their use in commercial products.



Conclusions

Overall, this study reveals a broad spectrum of maturity and clinical use of AI products, but a large gap exists in exploring the actual performance of AI tools in clinical practice.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{bucher2020digital,

title = {Digital health innovations, tools, and resources to support Helping Babies Survive programs},

author = {Sherri L. Bucher and Peter Cardellichio and Naomi Muinga and Jackie K. Patterson and Anu Thukral and Ashok K. Deorari and Santorino Data and Rachel Umoren and Saptarshi Purkayastha},

url = {https://pediatrics.aappublications.org/content/146/Supplement_2/S165},

doi = {10.1542/peds.2020-016915I},

issn = {0031-4005},

year  = {2020},

date = {2020-10-01},

journal = {Pediatrics},

volume = {146},

number = {Supplement 2},

pages = {S165-S182},

abstract = {The Helping Babies Survive (HBS) initiative features a suite of evidence-based curricula and simulation-based training programs designed to provide health workers in low- and middle-income countries (LMICs) with the knowledge, skills, and competencies to prevent, recognize, and manage leading causes of newborn morbidity and mortality. Global scale-up of HBS initiatives has been rapid. As HBS initiatives rolled out across LMIC settings, numerous bottlenecks, gaps, and barriers to the effective, consistent dissemination and implementation of the programs, across both the pre- and in-service continuums, emerged. Within the first decade of expansive scale-up of HBS programs, mobile phone ownership and access to cellular networks have also concomitantly surged in LMICs. In this article, we describe a number of HBS digital health innovations and resources that have been developed from 2010 to 2020 to support education and training, data collection for monitoring and evaluation, clinical decision support, and quality improvement. Helping Babies Survive partners and stakeholders can potentially integrate the described digital tools with HBS dissemination and implementation efforts in a myriad of ways to support low-dose high-frequency skills practice, in-person refresher courses, continuing medical and nursing education, on-the-job training, or peer-to-peer learning, and strengthen data collection for key newborn care and quality improvement indicators and outcomes. Thoughtful integration of purpose-built digital health tools, innovations, and resources may assist HBS practitioners to more effectively disseminate and implement newborn care programs in LMICs, and facilitate progress toward the achievement of Sustainable Development Goal health goals, targets, and objectives.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}