Informing clinical assessment by contextualizing post-hoc explanations of risk prediction models in type-2 diabetes.

Medical experts may use Artificial Intelligence (AI) systems with greater trust if these are supported by 'contextual explanations' that let the practitioner connect system inferences to their context of use. However, their importance in improving model usage and understanding has not been extensively studied. Hence, we consider a comorbidity risk prediction scenario and focus on contexts regarding the patients' clinical state, AI predictions about their risk of complications, and algorithmic explanations supporting the predictions. We explore how relevant information for such dimensions can be extracted from Medical guidelines to answer typical questions from clinical practitioners. We identify this as a question answering (QA) task and employ several state-of-the-art Large Language Models (LLM) to present contexts around risk prediction model inferences and evaluate their acceptability. Finally, we study the benefits of contextual explanations by building an end-to-end AI pipeline including data cohorting, AI risk modeling, post-hoc model explanations, and prototyped a visual dashboard to present the combined insights from different context dimensions and data sources, while predicting and identifying the drivers of risk of Chronic Kidney Disease (CKD) - a common type-2 diabetes (T2DM) comorbidity. All of these steps were performed in deep engagement with medical experts, including a final evaluation of the dashboard results by an expert medical panel. We show that LLMs, in particular BERT and SciBERT, can be readily deployed to extract some relevant explanations to support clinical usage. To understand the value-add of the contextual explanations, the expert panel evaluated these regarding actionable insights in the relevant clinical setting. Overall, our paper is one of the first end-to-end analyses identifying the feasibility and benefits of contextual explanations in a real-world clinical use case. Our findings can help improve clinicians' usage of AI models.

Quantifying representativeness in randomized clinical trials using machine learning fairness metrics.

OBJECTIVE: We help identify subpopulations underrepresented in randomized clinical trials (RCTs) cohorts with respect to national, community-based or health system target populations by formulating population representativeness of RCTs as a machine learning (ML) fairness problem, deriving new representation metrics, and deploying them in easy-to-understand interactive visualization tools. MATERIALS AND METHODS: We represent RCT cohort enrollment as random binary classification fairness problems, and then show how ML fairness metrics based on enrollment fraction can be efficiently calculated using easily computed rates of subpopulations in RCT cohorts and target populations. We propose standardized versions of these metrics and deploy them in an interactive tool to analyze 3 RCTs with respect to type 2 diabetes and hypertension target populations in the National Health and Nutrition Examination Survey. RESULTS: We demonstrate how the proposed metrics and associated statistics enable users to rapidly examine representativeness of all subpopulations in the RCT defined by a set of categorical traits (eg, gender, race, ethnicity, smoking status, and blood pressure) with respect to target populations. DISCUSSION: The normalized metrics provide an intuitive standardized scale for evaluating representation across subgroups, which may have vastly different enrollment fractions and rates in RCT study cohorts. The metrics are beneficial complements to other approaches (eg, enrollment fractions) used to identify generalizability and health equity of RCTs. CONCLUSION: By quantifying the gaps between RCT and target populations, the proposed methods can support generalizability evaluation of existing RCT cohorts. The interactive visualization tool can be readily applied to identified underrepresented subgroups with respect to any desired source or target populations.

Evaluating artificial intelligence in medicine: phases of clinical research.

Increased scrutiny of artificial intelligence (AI) applications in healthcare highlights the need for real-world evaluations for effectiveness and unintended consequences. The complexity of healthcare, compounded by the user- and context-dependent nature of AI applications, calls for a multifaceted approach beyond traditional in silico evaluation of AI. We propose an interdisciplinary, phased research framework for evaluation of AI implementations in healthcare. We draw analogies to and highlight differences from the clinical trial phases for drugs and medical devices, and we present study design and methodological guidance for each stage.

Knowledge Extraction of Cohort Characteristics in Research Publications.

When healthcare providers review the results of a clinical trial study to understand its applicability to their practice, they typically analyze how well the characteristics of the study cohort correspond to those of the patients they see. We have previously created a study cohort ontology to standardize this information and make it accessible for knowledge-based decision support. The extraction of this information from research publications is challenging, however, given the wide variance in reporting cohort characteristics in a tabular representation. To address this issue, we have developed an ontology-enabled knowledge extraction pipeline for automatically constructing knowledge graphs from the cohort characteristics found in PDF-formatted research papers. We evaluated our approach using a training and test set of 41 research publications and found an overall accuracy of 83.3% in correctly assembling the knowledge graphs. Our research provides a promising approach for extracting knowledge more broadly from tabular information in research publications.

A visual backchannel for large-scale events.

We introduce the concept of a Visual Backchannel as a novel way of following and exploring online conversations about large-scale events. Microblogging communities, such as Twitter, are increasingly used as digital backchannels for timely exchange of brief comments and impressions during political speeches, sport competitions, natural disasters, and other large events. Currently, shared updates are typically displayed in the form of a simple list, making it difficult to get an overview of the fast-paced discussions as it happens in the moment and how it evolves over time. In contrast, our Visual Backchannel design provides an evolving, interactive, and multi-faceted visual overview of large-scale ongoing conversations on Twitter. To visualize a continuously updating information stream, we include visual saliency for what is happening now and what has just happened, set in the context of the evolving conversation. As part of a fully web-based coordinated-view system we introduce Topic Streams, a temporally adjustable stacked graph visualizing topics over time, a People Spiral representing participants and their activity, and an Image Cloud encoding the popularity of event photos by size. Together with a post listing, these mutually linked views support cross-filtering along topics, participants, and time ranges. We discuss our design considerations, in particular with respect to evolving visualizations of dynamically changing data. Initial feedback indicates significant interest and suggests several unanticipated uses.