Saptarshi Purkayastha
Permanent URI for this collection
Risks and Opportunities of AI Recognition of Patient Race in Medical Imaging
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. His recent work published in Lancet Digital Health demonstrates that deep learning models have extremely high accuracy at identifying self-reported race from medical images such as X-rays, MRIs and CTs. This ability raises serious concerns among some researchers. Such software might group patients, or influence their care, by factoring in race. These AI models work very well on poor quality, distorted and even images where many parts of the image were deliberately cut out. These types of categorizations could lead to inequality in providing health care and making recommendations, and human decision makers might not understand how and why AI models are making the recommendations. Engineers, clinical researchers and informaticians need to get together to identify how AI models are able to have these superhuman capabilities.
Professor Purkayastha's translation of research into potential ways to identify and mitigate risks of deploying AI models in clinical practice to avoid racial issues in healthcare treatment is another example of how IUPUI's faculty members are TRANSLATING their RESEARCH INTO PRACTICE.
Browse
Recent Submissions
Item MedShift: identifying shift data for medical dataset curation(2021) Guo, Xiaoyuan; Gichoya, Judy Wawira; Trivedi, Hari; Purkayastha, Saptarshi; Banerjee, Imon; BioHealth Informatics, School of Informatics and ComputingTo curate a high-quality dataset, identifying data variance between the internal and external sources is a fundamental and crucial step. However, methods to detect shift or variance in data have not been significantly researched. Challenges to this are the lack of effective approaches to learn dense representation of a dataset and difficulties of sharing private data across medical institutions. To overcome the problems, we propose a unified pipeline called MedShift to detect the top-level shift samples and thus facilitate the medical curation. Given an internal dataset A as the base source, we first train anomaly detectors for each class of dataset A to learn internal distributions in an unsupervised way. Second, without exchanging data across sources, we run the trained anomaly detectors on an external dataset B for each class. The data samples with high anomaly scores are identified as shift data. To quantify the shiftness of the external dataset, we cluster B's data into groups class-wise based on the obtained scores. We then train a multi-class classifier on A and measure the shiftness with the classifier's performance variance on B by gradually dropping the group with the largest anomaly score for each class. Additionally, we adapt a dataset quality metric to help inspect the distribution differences for multiple medical sources. We verify the efficacy of MedShift with musculoskeletal radiographs (MURA) and chest X-rays datasets from more than one external source. Experiments show our proposed shift data detection pipeline can be beneficial for medical centers to curate high-quality datasets more efficiently. An interface introduction video to visualize our results is available at https://youtu.be/V3BF0P1sxQE.Item Human Activity Recognition using Deep Learning Models on Smartphones and Smartwatches Sensor Data(Scitepress, 2021) Oluwalade, Bolu; Neela, Sunil; Wawira, Judy; Adejumo, Tobiloba; Purkayastha, Saptarshi; BioHealth Informatics, School of Informatics and ComputingIn recent years, human activity recognition has garnered considerable attention both in industrial and academic research because of the wide deployment of sensors, such as accelerometers and gyroscopes, in products such as smartphones and smartwatches. Activity recognition is currently applied in various fields where valuable information about an individual’s functional ability and lifestyle is needed. In this study, we used the popular WISDM dataset for activity recognition. Using multivariate analysis of covariance (MANCOVA), we established a statistically significant difference (p < 0.05) between the data generated from the sensors embedded in smartphones and smartwatches. By doing this, we show that smartphones and smartwatches don’t capture data in the same way due to the location where they are worn. We deployed several neural network architectures to classify 15 different hand and non-hand oriented activities. These models include Long short-term memory (LSTM), Bi-directional Long short-term memory (BiLSTM), Convolutional Neural Network (CNN), and Convolutional LSTM (ConvLSTM). The developed models performed best with watch accelerometer data. Also, we saw that the classification precision obtained with the convolutional input classifiers (CNN and ConvLSTM) was higher than the end-to-end LSTM classifier in 12 of the 15 activities. Additionally, the CNN model for the watch accelerometer was better able to classify non-hand oriented activities when compared to hand-oriented activities.Item Big Data Analytics for developing countries – Using the Cloud for Operational BI in Health(Wiley, 2013) Braa, Jørn; Purkayastha, SaptarshiThe multi-layered view of digital divide suggests there is inequality of access to ICT, inequality of capability to exploit ICT and inequality of outcomes after exploiting ICT. This is evidently clear in the health systems of developing countries. In this paper, we look at cloud computing being able to provide computing as a utility service that might bridge this digital divide for Health Information Systems in developing countries. We highlight the role of Operational Business Intelligence (BI) tools to be able to make better decisions in health service provisioning. Through the case of DHIS2 software and its Analytics-as-a-Service (AaaS) model, we look at how tools can exploit Cloud computing capabilities to perform analytics on Big Data that is resulting from integration of health data from multiple sources. Beyond looking at purely warehousing techniques, we suggest understanding Big Data from Organizational Capabilities and expanding organizational capabilities by offloading computing as a utility to vendors through cloud computing.Item A Post-development Perspective on mHealth -- An Implementation Initiative in Malawi(IEEE, 2013-03-18) Purkayastha, Saptarshi; Manda, Tiwonge Davis; Sanner, Terje AkselWhile the sheer number of mHealth implementations around the world have been increasing dramatically, authoritative voices on global health have tried to put the focus on quantifiable evaluations and comparisons of these projects (e.g. health outcomes, cost savings, efficiency) in order to channel donor funds and investments into proven and scalable solutions. Drawing on empirical data from an mHealth implementation in Malawi we argue that quantitative evaluation of health interventions often assumes a top-down and limited view on the developmental impact of mHealth. Through our action-research involvement with facility-based reporting of routine health data through mobile phones, we conclude that developmental impacts of mHealth are local and each locale experience a different developmental impact depending on the context of use and available resources. The paper contrasts global concerns for quantifiable development with local priorities with respect to mHealth projects and information system (IS) interventions in health more broadly.Item Using cognitive fit theory to evaluate patient understanding of medical images(IEEE, 2017) Gichoya, Judy Wawira; Alarifi, Mohammad; Bhaduri, Ria; Tahir, Bilal; Purkayastha, Saptarshi; Radiology and Imaging Sciences, School of MedicinePatients are increasingly presented with their health data through patient portals in an attempt to engage patients in their own care. Due to the large amounts of data generated during a patient visit, the medical information when shared with patients can be overwhelming and cause anxiety due to lack of understanding. Health care organizations are attempting to improve transparency by providing patients with access to visit information. In this paper, we present our findings from a research study to evaluate patient understanding of medical images. We used cognitive fit theory to evaluate existing tools and images that are shared with patients and analyzed the relevance of such sharing. We discover that medical images need a lot of customization before they can be shared with patients. We suggest that new tools for medical imaging should be developed to fit the cognitive abilities of patients.Item Sustainable mobile information infrastructures in low resource settings(IOS, 2010) Braa, Kristin; Purkayastha, SaptarshiDeveloping countries represent the fastest growing mobile markets in the world. For people with no computing access, a mobile will be their first computing device. Mobile technologies offer a significant potential to strengthen health systems in developing countries with respect to community based monitoring, reporting, feedback to service providers, and strengthening communication and coordination between different health functionaries, medical officers and the community. However, there are various challenges in realizing this potential including technological such as lack of power, social, institutional and use issues. In this paper a case study from India on mobile health implementation and use will be reported. An underlying principle guiding this paper is to see mobile technology not as a "stand alone device" but potentially an integral component of an integrated mobile supported health information infrastructure.Item Exploring the potential and challenges of using mobile based technology in strengthening health information systems: Experiences from a pilot study(Association for Information Systems, 2010-08-01) Purkayastha, Saptarshi; Sahay, Sundeep; Mukherjee, ArunimaThis paper empirically examines the challenges of introducing a mobile based reporting system (called SCDRT) within the public health system in India to strengthen the health information systems, and also discusses the approaches to address these challenges. Taking an “infrastructure” perspective, various socio$technical challenges relating to technology, operator and usage are discussed. Scaling, in geographical and functional terms, is discussed with a focus on aspects of “attractors” and “motivation.”Item Implementation of a single sign-on system between practice, research and learning systems(Thieme, 2017-03-29) Purkayastha, Saptarshi; Gichoya, Judy W.; Addepally, Siva Abhishek; BioHealth Informatics, School of Informatics and ComputingBackground: Multiple specialized electronic medical systems are utilized in the health enterprise. Each of these systems has their own user management, authentication and authorization process, which makes it a complex web for navigation and use without a coherent process workflow. Users often have to remember multiple passwords, login/logout between systems that disrupt their clinical workflow. Challenges exist in managing permissions for various cadres of health care providers. Objectives: This case report describes our experience of implementing a single sign-on system, used between an electronic medical records system and a learning management system at a large academic institution with an informatics department responsible for student education and a medical school affiliated with a hospital system caring for patients and conducting research. Methods: At our institution, we use OpenMRS for research registry tracking of interventional radiology patients as well as to provide access to medical records to students studying health informatics. To provide authentication across different users of the system with different permissions, we developed a Central Authentication Service (CAS) module for OpenMRS, released under the Mozilla Public License and deployed it for single sign-on across the academic enterprise. The module has been in implementation since August 2015 to present, and we assessed usability of the registry and education system before and after implementation of the CAS module. 54 students and 3 researchers were interviewed. Results: The module authenticates users with appropriate privileges in the medical records system, providing secure access with minimal disruption to their workflow. No passwords requests were sent and users reported ease of use, with streamlined workflow. Conclusions: The project demonstrates that enterprise-wide single sign-on systems should be used in healthcare to reduce complexity like "password hell", improve usability and user navigation. We plan to extend this to work with other systems used in the health care enterprise.Item From Dyadic Ties to Information Infrastructures: Care-Coordination between Patients, Providers, Students and Researchers(Thieme, 2015-08-13) Purkayastha, Saptarshi; Price, A.; Biswas, R.; Jai Ganesh, A.U.; Otero, P.; BioHealth Informatics, School of Informatics and ComputingObjective: To share how an effectual merging of local and online networks in low resource regions can supplement and strengthen the local practice of patient centered care through the use of an online digital infrastructure powered by all stakeholders in healthcare. User Driven Health Care offers the dynamic integration of patient values and evidence based solutions for improved medical communication in medical care. Introduction: This paper conceptualizes patient care-coordination through the lens of engaged stakeholders using digital infrastructures tools to integrate information technology. We distinguish this lens from the prevalent conceptualization of dyadic ties between clinician-patient, patient-nurse, clinician-nurse, and offer the holistic integration of all stakeholder inputs, in the clinic and augmented by online communication in a multi-national setting. Methods: We analyze an instance of the user-driven health care (UDHC), a network of providers, patients, students and researchers working together to help manage patient care. The network currently focuses on patients from LMICs, but the provider network is global in reach. We describe UDHC and its opportunities and challenges in care-coordination to reduce costs, bring equity, and improve care quality and share evidence. Conclusion: UDHC has resulted in coordinated global based local care, affecting multiple facets of medical practice. Shared information resources between providers with disparate knowledge, results in better understanding by patients, unique and challenging cases for students, innovative community based research and discovery learning for all.Item Phronesis of AI in radiology: Superhuman meets natural stupidity(arXiv, 2018) Gichoya, Judy W.; Nuthakki, Siddhartha; Maity, Pallavi G.; Purkayastha, Saptarshi; BioHealth Informatics, School of Informatics and ComputingAdvances in AI in the last decade have clearly made economists, politicians, journalists, and citizenry in general believe that the machines are coming to take human jobs. We review 'superhuman' AI performance claims in radiology and then provide a self-reflection on our own work in the area in the form of a critical review, a tribute of sorts to McDermotts 1976 paper, asking the field for some self-discipline. Clearly there is an opportunity to replace humans, but there are better opportunities, as we have discovered to fit cognitive abilities of human and non-humans. We performed one of the first studies in radiology to see how human and AI performance can complement and improve each others performance for detecting pneumonia in chest X-rays. We question if there is a practical wisdom or phronesis that we need to demonstrate in AI today as well as in our field. Using this, we articulate what AI as a field has already and probably can in the future learn from Psychology, Cognitive Science, Sociology and Science and Technology Studies.