Towards Precision Medicine in Spinal Surgery: Leveraging AI Technologies

This critique explores the implications of integrating artificial intelligence (AI) technology, specifically OpenAI's advanced language model GPT-4 and its interface, ChatGPT, into the field of spinal surgery. It examines the potential effects of algorithmic bias, unique challenges in surgical domains, access and equity issues, cost implications, global disparities in technology adoption, and the concept of technological determinism. It posits that biases present in AI training data may impact the quality and equity of healthcare outcomes. Challenges related to the unique nature of surgical procedures, including real-time decision-making, are also addressed. Concerns over access, equity, and cost implications underscore the potential for exacerbated healthcare disparities. Global disparities in technology adoption highlight the importance of global collaboration, technology transfer, and capacity building. Finally, the critique challenges the notion of technological determinism, emphasizing the continued importance of human judgement and patient-care provider relationship in healthcare. The critique calls for a comprehensive evaluation of AI technology integration in healthcare to ensure equitable and quality care.

The effects of the COVID-19 pandemic for artificial intelligence practitioners: the decrease in tacit knowledge sharing

Purpose

This study aims to contribute by showing that although artificial intelligence (AI) practitioners have been faster to adapt, redefine and improve their remote working performance for routine tasks, they have instead decreased their tacit knowledge sharing and ability to perform extra tasks and manage the diverse time allocation.

Design/methodology/approach

Based on a grounded theory study of 57 in-depth interviews, conducted before the outbreak of the pandemic and after, this study investigates how remote work as a pandemic response measure affected AI practitioners.

Findings

Although remote working was a reality for AI practitioners before the COVID-19 pandemic, the overall remote working restrictions appear to have affected tacit knowledge sharing between AI practitioners, with a consequent negative impact on AI project output diversity.

Originality/value

The interactions of AI practitioners are partly embedded in AI tools and partly in human exchange. During the COVID-19 pandemic, these interactions appear to have become more obvious, even if the consequences have been unforeseen.

Health Decision Analysis: Evolution, Trends, and Emerging Topics

Health remains one of the most challenging realms for decision makers and policy making while critical for the well-being of humans, the stability of societies, and the development of economies. Decision making in this field ranges from medical doctors identifying the best treatments for patients, healthcare companies selecting the most promising drugs for development, healthcare providers deciding for adequate levels of resourcing, health regulators deciding whether to approve a new medicine or health technology, to regional and national health departments identifying how to increase the health security of regions and countries. In this positioning paper, and introduction to this Special Issue, we present the history, evolution, and trends of health decision analysis and suggest that these developments and news trends can be conceptualized as an emerging field of applied research for our discipline: Health Decision Analysis.

Information Systems Research for Smart Sustainable Mobility: A Framework and Call for Action

Transportation is a backbone of modern globalized societies. It also causes approximately one third of all European Union and U.S. greenhouse gas emissions, represents a major health hazard for global populations, and poses significant economic costs. However, rapid innovation in vehicle technology, mobile connectivity, computing hardware, and artificial intelligence (AI)-powered information systems heralds a deep socio-technical transformation of the sector. The emergence of connected, autonomous, shared, and electric (CASE) vehicle technology has created a digital layer that resides on top of the traditional physical mobility system. This article contributes a framework to direct research and practice toward leveraging the opportunities afforded by CASE for a more efficient and less environmentally problematic mobility system. The authors propose seven overarching dimensions of action. These range from designing real-time digital coordination mechanisms for the management of mobility systems to developing AI-powered real-time decision support for mobility resource planning and operations. Per each dimension, concrete angles of attack are suggested which, we hope, will spur structured engagement from both researchers and practitioners in the field.

Integrating Behavioral, Economic, and Technical Insights to Understand and Address Algorithmic Bias: A Human-Centric Perspective

Many important decisions are increasingly being made with the help of information systems that use artificial intelligence and machine learning models. These computational models are designed to discover useful patterns from large amounts of data, which augment human capabilities to make decisions in various application domains. However, there are growing concerns regarding the ethics challenges faced by these automated decision-making (ADM) models, most notably on the issue of algorithmic bias , where the models systematically produce less favorable (i.e., unfair) decisions for certain groups of people. In this commentary, we argue that algorithmic bias is not just a technical (e.g., computational or statistical) problem, and its successful resolution requires deep insights into individual and organizational behavior, economic incentives, as well as complex dynamics of the sociotechnical systems in which the ADM models are embedded. We discuss a human-centric, fairness-aware ADM framework which highlights the holistic involvement of human decision makers in each step of ADM. We review the emerging literature on fairness-aware machine learning, and then discuss various strategic decisions that humans need to make, such as formulating proper fairness objectives, recognizing fairness-induced trade-offs and implications, utilizing machine learning model outputs, and managing/governing the decisions of ADM models. We further illustrate how these strategic decisions are jointly informed by behavioral, economic, and design sciences. Our discussions reveal a number of future research opportunities uniquely suitable for Management Information Systems (MIS) researchers to pursue.

An Adaptive Deep Ensemble Learning Method for Dynamic Evolving Diagnostic Task Scenarios

Increasingly, machine learning methods have been applied to aid in diagnosis with good results. However, some complex models can confuse physicians because they are difficult to understand, while data differences across diagnostic tasks and institutions can cause model performance fluctuations. To address this challenge, we combined the Deep Ensemble Model (DEM) and tree-structured Parzen Estimator (TPE) and proposed an adaptive deep ensemble learning method (TPE-DEM) for dynamic evolving diagnostic task scenarios. Different from previous research that focuses on achieving better performance with a fixed structure model, our proposed model uses TPE to efficiently aggregate simple models more easily understood by physicians and require less training data. In addition, our proposed model can choose the optimal number of layers for the model and the type and number of basic learners to achieve the best performance in different diagnostic task scenarios based on the data distribution and characteristics of the current diagnostic task. We tested our model on one dataset constructed with a partner hospital and five UCI public datasets with different characteristics and volumes based on various diagnostic tasks. Our performance evaluation results show that our proposed model outperforms other baseline models on different datasets. Our study provides a novel approach for simple and understandable machine learning models in tasks with variable datasets and feature sets, and the findings have important implications for the application of machine learning models in computer-aided diagnosis.

Augmenting Medical Diagnosis Decisions? An Investigation into Physicians’ Decision-Making Process with Artificial Intelligence

Systems based on artificial intelligence (AI) increasingly support physicians in diagnostic decisions, but they are not without errors and biases. Failure to detect those may result in wrong diagnoses and medical errors. Compared with rule-based systems, however, these systems are less transparent and their errors less predictable. Thus, it is difficult, yet critical, for physicians to carefully evaluate AI advice. This study uncovers the cognitive challenges that medical decision makers face when they receive potentially incorrect advice from AI-based diagnosis systems and must decide whether to follow or reject it. In experiments with 68 novice and 12 experienced physicians, novice physicians with and without clinical experience as well as experienced radiologists made more inaccurate diagnosis decisions when provided with incorrect AI advice than without advice at all. We elicit five decision-making patterns and show that wrong diagnostic decisions often result from shortcomings in utilizing metacognitions related to decision makers’ own reasoning (self-monitoring) and metacognitions related to the AI-based system (system monitoring). As a result, physicians fall for decisions based on beliefs rather than actual data or engage in unsuitably superficial evaluation of the AI advice. Our study has implications for the training of physicians and spotlights the crucial role of human actors in compensating for AI errors.

Designing Payment Contracts for Healthcare Services to Induce Information Sharing: The Adoption and the Value of Health Information Exchanges (HIEs)

Recent initiatives to improve healthcare quality and reduce costs have centered around payment mechanisms and IT-enabled health information exchanges (HIEs). Such initiatives profoundly influence both providers’ choices in terms of healthcare effort levels and HIE adoption and patients’ choice of providers. Using a game-theoretical model of a healthcare setup, we examine the role of payment models in aligning providers’ and patients’ incentives for realizing socially optimal (i.e., first-best) choices. We show that the traditional fee-for-service (FFS) payment model does not necessarily induce the first-best solution. The pay-for-performance (P4P) model may induce the first-best solution under some conditions if provider switching by patients during a health episode is socially suboptimal, making provider coordination less of an issue. We identify an episode-based payment (EBP) model that can always induce the first-best solution. The proposed EBP model reduces to the P4P model if the P4P model induces the first-best solution. In other cases, the first-best inducing EBP model is multilateral in the sense that the payment to a provider depends not only on the provider’s own efforts and outcomes but also on those of other providers. Furthermore, the payment in this EBP model is sequence dependent in the sense that payment to a provider is contingent upon whether the patient visits a given provider first or second. We show that the proposed EBP model achieves the lowest healthcare cost, not necessarily at the expense of care quality or provider payment, relative to FFS and P4P. Although our proposed contract is complex, it sets an optimality baseline when evaluating simpler contracts and also characterizes aspects of payment that need to be captured for socially desirable actions. We further show that the value of HIEs depends critically on the payment model as well as on the social desirability of patient switching. Under all three payment models, the HIE value is higher when switching by at least some patients is desirable than when switching by any patient is undesirable. Moreover, the HIE value is highest under the FFS model and lowest under the P4P model. Hence, assessing the value of HIEs in isolation from the underlying payment mechanism and patient-switching behavior may result in under- or overestimation of the HIE value. Therefore, as payment models evolve over time, there is a real need to reevaluate the HIE value and the government subsidies that induce providers to adopt HIEs.