Toward a responsible future: recommendations for AI-enabled clinical decision support.

BACKGROUND: Integrating artificial intelligence (AI) in healthcare settings has the potential to benefit clinical decision-making. Addressing challenges such as ensuring trustworthiness, mitigating bias, and maintaining safety is paramount. The lack of established methodologies for pre- and post-deployment evaluation of AI tools regarding crucial attributes such as transparency, performance monitoring, and adverse event reporting makes this situation challenging. OBJECTIVES: This paper aims to make practical suggestions for creating methods, rules, and guidelines to ensure that the development, testing, supervision, and use of AI in clinical decision support (CDS) systems are done well and safely for patients. MATERIALS AND METHODS: In May 2023, the Division of Clinical Informatics at Beth Israel Deaconess Medical Center and the American Medical Informatics Association co-sponsored a working group on AI in healthcare. In August 2023, there were 4 webinars on AI topics and a 2-day workshop in September 2023 for consensus-building. The event included over 200 industry stakeholders, including clinicians, software developers, academics, ethicists, attorneys, government policy experts, scientists, and patients. The goal was to identify challenges associated with the trusted use of AI-enabled CDS in medical practice. Key issues were identified, and solutions were proposed through qualitative analysis and a 4-month iterative consensus process. RESULTS: Our work culminated in several key recommendations: (1) building safe and trustworthy systems; (2) developing validation, verification, and certification processes for AI-CDS systems; (3) providing a means of safety monitoring and reporting at the national level; and (4) ensuring that appropriate documentation and end-user training are provided. DISCUSSION: AI-enabled Clinical Decision Support (AI-CDS) systems promise to revolutionize healthcare decision-making, necessitating a comprehensive framework for their development, implementation, and regulation that emphasizes trustworthiness, transparency, and safety. This framework encompasses various aspects including model training, explainability, validation, certification, monitoring, and continuous evaluation, while also addressing challenges such as data privacy, fairness, and the need for regulatory oversight to ensure responsible integration of AI into clinical workflow. CONCLUSIONS: Achieving responsible AI-CDS systems requires a collective effort from many healthcare stakeholders. This involves implementing robust safety, monitoring, and transparency measures while fostering innovation. Future steps include testing and piloting proposed trust mechanisms, such as safety reporting protocols, and establishing best practice guidelines.

Applying diagnosis support systems in electronic health records to identify wild-type transthyretin amyloid cardiomyopathy risk.

Aim: Wild-type transthyretin amyloid cardiomyopathy (ATTRwt-CM) is frequently misdiagnosed, and delayed diagnosis is associated with substantial morbidity and mortality. At three large academic medical centers, combinations of phenotypic features were implemented in electronic health record (EHR) systems to identify patients with heart failure at risk for ATTRwt-CM. Methods: Phenotypes/phenotype combinations were selected based on strength of correlation with ATTRwt-CM versus non-amyloid heart failure; different clinical decision support and reporting approaches and data sources were evaluated on Cerner and Epic EHR platforms. Results: Multiple approaches/sources showed potential usefulness for incorporating predictive analytics into the EHR to identify at-risk patients. Conclusion: These preliminary findings may guide other medical centers in building and implementing similar systems to improve recognition of ATTRwt-CM in patients with heart failure.

Usability and Perceived Usefulness of the AFib 2gether Mobile App in a Clinical Setting: Single-Arm Intervention Study.

BACKGROUND: Although the American Heart Association and other professional societies have recommended shared decision-making as a way for patients with atrial fibrillation (AF) or atrial flutter to make informed decisions about using anticoagulation (AC), the best method for facilitating shared decision-making remains uncertain. OBJECTIVE: The aim of this study is to assess the AFib 2gether mobile app for usability, perceived usefulness, and the extent and nature of shared decision-making that occurred for clinical encounters between patients with AF and their cardiology providers in which the app was used. METHODS: We identified patients visiting a cardiology provider between October 2019 and May 2020. We measured usability from patients and providers using the Mobile App Rating Scale. From the 8 items of the Mobile App Rating Scale, we reported the average score (out of 5) for domains of functionality, esthetics, and overall quality. We administered a 3-item questionnaire to patients relating to their perceived usefulness of the app and a separate 3-item questionnaire to providers to measure their perceived usefulness of the app. We performed a chart review to track the occurrence of AC within 6 months of the index visit. We also audio recorded a subset of the encounters to identify evidence of shared decision-making. RESULTS: We facilitated shared decision-making visits for 37 patients visiting 13 providers. In terms of usability, patients' average ratings of functionality, esthetics, and overall quality were 4.51 (SD 0.61), 4.26 (SD 0.51), and 4.24 (SD 0.89), respectively. In terms of usefulness, 41% (15/37) of patients agreed that the app improved their knowledge regarding AC, and 62% (23/37) agreed that the app helped clarify to their provider their preferences regarding AC. Among providers, 79% (27/34) agreed that the app helped clarify their patients' preferences, 82% (28/34) agreed that the app saved them time, and 59% (20/34) agreed that the app helped their patients make decisions about AC. In addition, 32% (12/37) of patients started AC after their shared decision-making visits. We audio recorded 25 encounters. Of these, 84% (21/25) included the mention of AC for AF, 44% (11/25) included the discussion of multiple options for AC, 72% (18/25) included a provider recommendation for AC, and 48% (12/25) included the evidence of patient involvement in the discussion. CONCLUSIONS: Patients and providers rated the app with high usability and perceived usefulness. Moreover, one-third of the patients began AC, and approximately 50% (12/25) of the encounters showed evidence of patient involvement in decision-making. In the future, we plan to study the effect of the app on a larger sample and with a controlled study design. TRIAL REGISTRATION: ClinicalTrials.gov NCT04118270; https://clinicaltrials.gov/ct2/show/NCT04118270. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): RR2-21986.

Using Artificial Intelligence With Natural Language Processing to Combine Electronic Health Record's Structured and Free Text Data to Identify Nonvalvular Atrial Fibrillation to Decrease Strokes and Death: Evaluation and Case-Control Study.

BACKGROUND: Nonvalvular atrial fibrillation (NVAF) affects almost 6 million Americans and is a major contributor to stroke but is significantly undiagnosed and undertreated despite explicit guidelines for oral anticoagulation. OBJECTIVE: The aim of this study is to investigate whether the use of semisupervised natural language processing (NLP) of electronic health record's (EHR) free-text information combined with structured EHR data improves NVAF discovery and treatment and perhaps offers a method to prevent thousands of deaths and save billions of dollars. METHODS: We abstracted 96,681 participants from the University of Buffalo faculty practice's EHR. NLP was used to index the notes and compare the ability to identify NVAF, congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, stroke or transient ischemic attack, vascular disease, age 65 to 74 years, sex category (CHA2DS2-VASc), and Hypertension, Abnormal liver/renal function, Stroke history, Bleeding history or predisposition, Labile INR, Elderly, Drug/alcohol usage (HAS-BLED) scores using unstructured data (International Classification of Diseases codes) versus structured and unstructured data from clinical notes. In addition, we analyzed data from 63,296,120 participants in the Optum and Truven databases to determine the NVAF frequency, rates of CHA2DS2­VASc ≥2, and no contraindications to oral anticoagulants, rates of stroke and death in the untreated population, and first year's costs after stroke. RESULTS: The structured-plus-unstructured method would have identified 3,976,056 additional true NVAF cases (P<.001) and improved sensitivity for CHA2DS2-VASc and HAS-BLED scores compared with the structured data alone (P=.002 and P<.001, respectively), causing a 32.1% improvement. For the United States, this method would prevent an estimated 176,537 strokes, save 10,575 lives, and save >US $13.5 billion. CONCLUSIONS: Artificial intelligence-informed bio-surveillance combining NLP of free-text information with structured EHR data improves data completeness, prevents thousands of strokes, and saves lives and funds. This method is applicable to many disorders with profound public health consequences.