Meeting the Moment: Addressing Barriers and Facilitating Clinical Adoption of Artificial Intelligence in Medical Diagnosis

Inherited retinal disease pathway in the UK: a patient perspective and the potential of AI

BACKGROUND

Inherited retinal diseases (IRDs) are the leading cause of blindness in young people in the UK. Despite significant improvements in genomics medicine, the diagnosis of these conditions remains challenging, and around 40% do not receive a definite genetic diagnosis after extensive genetic testing. This survey aims to investigate the experience of individuals affected by IRDs, their relatives, friends and caregivers, focusing on their care and diagnostic journey. Additionally, it explores the potential acceptability of artificial intelligence (AI) technologies, such as Eye2Gene, that predict causative genes from retinal images of patients with IRDs.

METHODS

This cross-sectional survey included Likert scale and open-ended questions and was distributed electronically using the Qualtrics platform between April and August 2024. The survey included questions on respondent demographics; their journey to receive specialist care and genetic testing; their information needs and their attitude towards AI-augmented diagnosis. Descriptive statistics and content analysis were used to interpret the survey responses.

RESULTS

The survey had 247 responses, of which 242 were analysed after removing four duplicates and one without consent; 80.2% were patients and the remainder were relatives, friends or caregivers. There was substantial variability in patient diagnostic journeys in terms of waiting times to see a specialist (IQR, 1-4 years), commute required (IQR, 10-74 miles) and number of visits to reach a diagnosis (IQR, 2-4). A substantial proportion of patients (35.8%) had a change in diagnosis. The majority of respondents (>90%) were overwhelmingly in favour of the integration of AI into the IRD pathway to accelerate genetic diagnosis and improve care.

CONCLUSION

This survey identifies several key gaps and disparities in the IRD care pathway which may potentially be bridged with AI. The survey also reveals a favourable attitude towards incorporating AI into diagnostic testing of IRDs.

Leveraging artificial intelligence to reduce diagnostic errors in emergency medicine: Challenges, opportunities, and future directions

Diagnostic errors in health care pose significant risks to patient safety and are disturbingly common. In the emergency department (ED), the chaotic and high-pressure environment increases the likelihood of these errors, as emergency clinicians must make rapid decisions with limited information, often under cognitive overload. Artificial intelligence (AI) offers promising solutions to improve diagnostic errors in three key areas: information gathering, clinical decision support (CDS), and feedback through quality improvement. AI can streamline the information-gathering process by automating data retrieval, reducing cognitive load, and providing clinicians with essential patient details quickly. AI-driven CDS systems enhance diagnostic decision making by offering real-time insights, reducing cognitive biases, and prioritizing differential diagnoses. Furthermore, AI-powered feedback loops can facilitate continuous learning and refinement of diagnostic processes by providing targeted education and outcome feedback to clinicians. By integrating AI into these areas, the potential for reducing diagnostic errors and improving patient safety in the ED is substantial. However, successfully implementing AI in the ED is challenging and complex. Developing, validating, and implementing AI as a safe, human-centered ED tool requires thoughtful design and meticulous attention to ethical and practical considerations. Clinicians and patients must be integrated as key stakeholders across these processes. Ultimately, AI should be seen as a tool that assists clinicians by supporting better, faster decisions and thus enhances patient outcomes.

Enhancing medical imaging education: integrating computing technologies, digital image processing and artificial intelligence

The rapid advancement of technology has brought significant changes to various fields, including medical imaging (MI). This discussion paper explores the integration of computing technologies (e.g. Python and MATLAB), digital image processing (e.g. image enhancement, segmentation and three-dimensional reconstruction) and artificial intelligence (AI) into the undergraduate MI curriculum. By examining current educational practices, gaps and limitations that hinder the development of future-ready MI professionals are identified. A comprehensive curriculum framework is proposed, incorporating essential computational skills, advanced image processing techniques and state-of-the-art AI tools, such as large language models like ChatGPT. The proposed curriculum framework aims to improve the quality of MI education significantly and better equip students for future professional practice and challenges while enhancing diagnostic accuracy, improving workflow efficiency and preparing students for the evolving demands of the MI field.

Reporting Quality of AI Intervention in Randomized Controlled Trials in Primary Care: Systematic Review and Meta-Epidemiological Study

BACKGROUND

The surge in artificial intelligence (AI) interventions in primary care trials lacks a study on reporting quality.

OBJECTIVE

This study aimed to systematically evaluate the reporting quality of both published randomized controlled trials (RCTs) and protocols for RCTs that investigated AI interventions in primary care.

METHODS

PubMed, Embase, Cochrane Library, MEDLINE, Web of Science, and CINAHL databases were searched for RCTs and protocols on AI interventions in primary care until November 2024. Eligible studies were published RCTs or full protocols for RCTs exploring AI interventions in primary care. The reporting quality was assessed using CONSORT-AI (Consolidated Standards of Reporting Trials-Artificial Intelligence) and SPIRIT-AI (Standard Protocol Items: Recommendations for Interventional Trials-Artificial Intelligence) checklists, focusing on AI intervention-related items.

RESULTS

A total of 11,711 records were identified. In total, 19 published RCTs and 21 RCT protocols for 35 trials were included. The overall proportion of adequately reported items was 65% (172/266; 95% CI 59%-70%) and 68% (214/315; 95% CI 62%-73%) for RCTs and protocols, respectively. The percentage of RCTs and protocols that reported a specific item ranged from 11% (2/19) to 100% (19/19) and from 10% (2/21) to 100% (21/21), respectively. The reporting of both RCTs and protocols exhibited similar characteristics and trends. They both lack transparency and completeness, which can be summarized in three aspects: without providing adequate information regarding the input data, without mentioning the methods for identifying and analyzing performance errors, and without stating whether and how the AI intervention and its code can be accessed.

CONCLUSIONS

The reporting quality could be improved in both RCTs and protocols. This study helps promote the transparent and complete reporting of trials with AI interventions in primary care.

Revolutionizing diabetic retinopathy screening and management: The role of artificial intelligence and machine learning

Diabetic retinopathy (DR) remains a leading cause of vision impairment and blindness among individuals with diabetes, necessitating innovative approaches to screening and management. This editorial explores the transformative potential of artificial intelligence (AI) and machine learning (ML) in revolutionizing DR care. AI and ML technologies have demonstrated remarkable advancements in enhancing the accuracy, efficiency, and accessibility of DR screening, helping to overcome barriers to early detection. These technologies leverage vast datasets to identify patterns and predict disease progression with unprecedented precision, enabling clinicians to make more informed decisions. Furthermore, AI-driven solutions hold promise in personalizing management strategies for DR, incorporating predictive analytics to tailor interventions and optimize treatment pathways. By automating routine tasks, AI can reduce the burden on healthcare providers, allowing for a more focused allocation of resources towards complex patient care. This review aims to evaluate the current advancements and applications of AI and ML in DR screening, and to discuss the potential of these technologies in developing personalized management strategies, ultimately aiming to improve patient outcomes and reduce the global burden of DR. The integration of AI and ML in DR care represents a paradigm shift, offering a glimpse into the future of ophthalmic healthcare.

Hermeneutics as impediment to AI in medicine

Predictions that artificial intelligence (AI) will become capable of replacing human beings in domains such as medicine rest implicitly on a theory of mind according to which knowledge can be captured propositionally without loss of meaning. Generative AIs, for example, draw upon billions of written sources to produce text that most likely responds to a user's query, according to its probability heuristic. Such programs can only replace human beings in practices such as medicine if human language functions similarly and, like AI, does not rely on meta-textual resources to convey meaning. In this essay, I draw on the hermeneutic philosophy of Hans-Georg Gadamer to challenge this conception of human knowledge. I follow Gadamer in arguing that human understanding of texts is an interpretive process relying on previously received judgments that derive from the human person's situatedness in history, and these judgments differ from the rules guiding generative AI. Human understanding is also dialogical, as it depends on the 'fusion of horizons' with another person to the extent that one's own prejudices may come under question, something AI cannot achieve. Furthermore, artificial intelligence lacks a human body, which conditions human perception and understanding. I contend that these non-textual sources of meaning, which must remain obscure to AI, are important in moral practices such as medicine, particularly in history-taking, physical examination, diagnostic reasoning, and negotiating a treatment plan. Although AI can undoubtedly aid physicians in certain ways, it faces inherent limitations in replicating these core tasks of the physician-patient relationship.

Assessing the disconnect between student interest and education in artificial intelligence in medicine in Saudi Arabia

BACKGROUND

Although artificial intelligence (AI) has gained increasing attention for its potential future impact on clinical practice, medical education has struggled to stay ahead of the developing technology. The question of whether medical education is fully preparing trainees to adapt to potential changes from AI technology in clinical practice remains unanswered, and the influence of AI on medical students' career preferences remains unclear. Understanding the gap between students' interest in and knowledge of AI may help inform the medical curriculum structure.

METHODS

A total of 354 medical students were surveyed to investigate their knowledge of, exposure to, and interest in the role of AI in health care. Students were questioned about the anticipated impact of AI on medical specialties and their career preferences.

RESULTS

Most students (65%) were interested in the role of AI in medicine, but only 23% had received formal education in AI based on reliable scientific resources. Despite their interest and willingness to learn, only 20.1% of students reported that their school offered resources enabling them to explore the use of AI in medicine. They relied mainly on informal information sources, including social media, and few students understood fundamental AI concepts or could cite clinically relevant AI research. Students who cited more scientific primary sources (rather than online media) exhibited significantly higher self-reported understanding of AI concepts in the context of medicine. Interestingly, students who had received more exposure to AI courses reported higher levels of skepticism regarding AI and were less eager to learn more about it. Radiology and pathology were perceived to be the fields most strongly affected by AI. Students reported that their overall choice of specialty was not impacted by AI.

CONCLUSION

Formal AI education seems inadequate despite students' enthusiasm concerning the application of such technology in clinical practice. Medical curricula should evolve to promote structured, evidence-based AI literacy to enable students to understand the potential applications of AI in health care.

AI Code of Conduct-Safety, Inclusivity, and Sustainability

This Viewpoint discusses the need for guidance and governance in assessing the safety of health care artificial intelligence (AI) and machine learning systems.

The Role of Machine Learning Models in Predicting Cirrhosis Mortality: A Systematic Review

Liver cirrhosis affects millions of individuals worldwide and is one of the primary causes of mortality. Early mortality prediction for cirrhosis patients may increase the possibility for medical professionals to treat the illness successfully. This study assesses the ability of machine learning (ML) models to predict cirrhosis mortality. We followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to search for relevant literature across four different databases. We found 379 studies of which 10 were eligible for inclusion in the current study. We analyzed 10 retrospective studies that showed that ML models outperformed conventional scores in predicting the death rate from end-stage liver disease (ESLD). Interestingly, models that used more parameters, such as patient demographics and extensive laboratory testing, exhibited higher prediction accuracy. With an area under the receiver operating characteristic (AUROC) ranging from 0.71 to 0.96, ML models showed consistently significant gains over traditional prognostic ratings. This review emphasizes how ML models might improve ESLD patient death prediction. Because machine learning models are more accurate than conventional approaches, it is important to incorporate data-driven informatics technologies into clinical settings. Additional validation and openness are required to guarantee model dependability and interpretability before ML may be used in clinical practice. The goal of future research should be to create reliable, interpretable models that may be used successfully in a variety of clinical contexts, enhancing ESLD patient treatment and results.

Harnessing artificial intelligence for enhanced veterinary diagnostics: A look to quality assurance, Part I Model development

Artificial intelligence (AI) has transformative potential in veterinary pathology in tasks ranging from cell enumeration and cancer detection to prognosis forecasting, virtual staining techniques, and individually tailored treatment plans. Preclinical testing and validation of AI systems (AIS) are critical to ensure diagnostic safety, efficacy, and dependability. In this two-part series, challenges such as the AI chasm (ie, the discrepancy between the AIS model performance in research settings and real-world applications) and ethical considerations (data privacy, algorithmic bias) are reviewed and underscore the importance of tailored quality assurance measures that address the nuances of AI in veterinary pathology. This review advocates for a multidisciplinary approach to AI development and implementation, focusing on image-based tasks, highlighting the necessity for collaboration across veterinarians, computer scientists, and ethicists to successfully navigate the complex landscape of using AI in veterinary medicine. It calls for a concerted effort to bridge the AI chasm by addressing technical, ethical, and regulatory challenges, facilitating AI integration into veterinary pathology. The future of veterinary pathology must balance harnessing AI's potential while intentionally mitigating its risks, ensuring the welfare of animals and the integrity of the veterinary profession are safeguarded. Part I of this review focuses on considerations for model development, and Part II focuses on external validation of AI.

An Interoperable Machine Learning Pipeline for Pediatric Obesity Risk Estimation

Reliable prediction of pediatric obesity can offer a valuable resource to providers, helping them engage in timely preventive interventions before the disease is established. Many efforts have been made to develop ML-based predictive models of obesity, and some studies have reported high predictive performances. However, no commonly used clinical decision support tool based on existing ML models currently exists. This study presents a novel end-to-end pipeline specifically designed for pediatric obesity prediction, which supports the entire process of data extraction, inference, and communication via an API or a user interface. While focusing only on routinely recorded data in pediatric electronic health records (EHRs), our pipeline uses a diverse expert-curated list of medical concepts to predict the 1-3 years risk of developing obesity. Furthermore, by using the Fast Healthcare Interoperability Resources (FHIR) standard in our design procedure, we specifically target facilitating low-effort integration of our pipeline with different EHR systems. In our experiments, we report the effectiveness of the predictive model as well as its alignment with the feedback from various stakeholders, including ML scientists, providers, health IT personnel, health administration representatives, and patient group representatives.

Using the Constrained Disorder Principle to Navigate Uncertainties in Biology and Medicine: Refining Fuzzy Algorithms

Uncertainty in biology refers to situations in which information is imperfect or unknown. Variability, on the other hand, is measured by the frequency distribution of observed data. Biological variability adds to the uncertainty. The Constrained Disorder Principle (CDP) defines all systems in the universe by their inherent variability. According to the CDP, systems exhibit a degree of variability necessary for their proper function, allowing them to adapt to changes in their environments. Per the CDP, while variability differs from uncertainty, it can be viewed as a regulated mechanism for efficient functionality rather than uncertainty. This paper explores the various aspects of un-certainties in biology. It focuses on using CDP-based platforms for refining fuzzy algorithms to address some of the challenges associated with biological and medical uncertainties. Developing a fuzzy decision tree that considers the natural variability of systems can help minimize uncertainty. This method can reveal previously unidentified classes, reduce the number of unknowns, improve the accuracy of modeling results, and generate algorithm outputs that are more biologically and clinically relevant.

Biomarker discovery in hepatocellular carcinoma (HCC) for personalized treatment and enhanced prognosis

Hepatocellular carcinoma (HCC) is a leading contributor to cancer-related deaths worldwide and presents significant challenges in diagnosis and treatment due to its heterogeneous nature. The discovery of biomarkers has become crucial in addressing these challenges, promising early detection, precise diagnosis, and personalized treatment plans. Key biomarkers, such as alpha fetoprotein (AFP) glypican 3 (GPC3) and des gamma carboxy prothrombin (DCP) have shown potential in improving clinical results. Progress in proteomic technologies, including next-generation sequencing (NGS), mass spectrometry, and liquid biopsies detecting circulating tumor cells (CTCs) and circulating tumor DNA (ctDNA), has deepened our understanding of HCC's molecular landscape. Immunological markers, like PD-L1 expression and tumor-infiltrating lymphocytes (TILs), also play a crucial role in guiding immunotherapy decisions. Despite these advancements, challenges remain in biomarker validation, standardization, integration into clinical practice, and cost-related barriers. Emerging technologies like single-cell sequencing and machine learning offer promising avenues for further exploration. Continued investment in research and collaboration among researchers, healthcare providers, and policymakers is vital to harness the potential of biomarkers fully, ultimately revolutionizing HCC management and improving patient outcomes through personalized treatment approaches.

Artificial Intelligence in the Provision of Health Care: An American College of Physicians Policy Position Paper

Internal medicine physicians are increasingly interacting with systems that implement artificial intelligence (AI) and machine learning (ML) technologies. Some physicians and health care systems are even developing their own AI models, both within and outside of electronic health record (EHR) systems. These technologies have various applications throughout the provision of health care, such as clinical documentation, diagnostic image processing, and clinical decision support. With the growing availability of vast amounts of patient data and unprecedented levels of clinician burnout, the proliferation of these technologies is cautiously welcomed by some physicians. Others think it presents challenges to the patient-physician relationship and the professional integrity of physicians. These dispositions are understandable, given the "black box" nature of some AI models, for which specifications and development methods can be closely guarded or proprietary, along with the relative lagging or absence of appropriate regulatory scrutiny and validation. This American College of Physicians (ACP) position paper describes the College's foundational positions and recommendations regarding the use of AI- and ML-enabled tools and systems in the provision of health care. Many of the College's positions and recommendations, such as those related to patient-centeredness, privacy, and transparency, are founded on principles in the ACP Ethics Manual. They are also derived from considerations for the clinical safety and effectiveness of the tools as well as their potential consequences regarding health disparities. The College calls for more research on the clinical and ethical implications of these technologies and their effects on patient health and well-being.

Kairos in diagnostics

Kairos has been a key concept in medicine for millennia and is frequently understood as "the right time" in relation to treatment. In this study we scrutinize kairos in the context of diagnostics. This has become highly topical as technological developments have caused diagnostics to be performed ever earlier in the disease development. Detecting risk factors, precursors, and predictors of disease (in biomarkers, pre-disease, and pre-pre-disease) has resulted in too early diagnoses, i.e., overdiagnoses. Nonetheless, despite vast advances in science and technology, diagnoses also come too late. Accordingly, timing diagnostics right is crucial. In this article we start with giving a brief overview of the etymology and general use of the concepts of kairos and diagnosis. Then we delimit kairos in diagnostics by analysing "too early" and "too late" diagnosis and by scrutinizing various phases of diagnostics. This leads us to define kairos of diagnostics as the time when there is potential for sufficient information for making a diagnosis that is most helpful for the person. It allows us to conclude that kairos is as important in diagnostics as in therapeutics.

Insights Into Vestibulo-Ocular Reflex Artifacts: A Narrative Review of the Video Head Impulse Test (vHIT)

Video head impulse test (vHIT) artifacts are defined as spurious elements or disturbances in the recorded data that deviate from the true vestibulo-ocular reflex response. These artifacts can arise from various sources, encompassing technological limitations, patient-specific factors, or environmental influences, introducing inaccuracies in vHIT outcomes. The absence of standardized criteria for artifact identification leads to methodological heterogeneity. This narrative review aims to comprehensively examine the challenges posed by artifacts in the vHIT. By surveying existing literature, the review seeks to elucidate the multifaceted nature of artifacts arising from technological, patient-related, evaluator-related, and environmental factors.

Achieving High Accuracy in Predicting the Probability of Periprosthetic Joint Infection From Synovial Fluid in Patients Undergoing Hip or Knee Arthroplasty: The Development and Validation of a Multivariable Machine Learning Algorithm

Background and objective The current periprosthetic joint infection (PJI) diagnostic guidelines require clinicians to interpret and integrate multiple criteria into a complex scoring system. Also, PJI classifications are often inconclusive, failing to provide a clinical diagnosis. Machine learning (ML) models could be leveraged to reduce reliance on these complex systems and thereby reduce diagnostic uncertainty. This study aimed to develop an ML algorithm using synovial fluid (SF) test results to establish a PJI probability score. Methods We used a large clinical laboratory's dataset of SF samples, aspirated from patients with hip or knee arthroplasty as part of a PJI evaluation. Patient age and SF biomarkers [white blood cell count, neutrophil percentage (%PMN), red blood cell count, absorbance at 280 nm wavelength, C-reactive protein (CRP), alpha-defensin (AD), neutrophil elastase, and microbial antigen (MID) tests] were used for model development. Data preprocessing, principal component analysis, and unsupervised clustering (K-means) revealed four clusters of samples that naturally aggregated based on biomarker results. Analysis of the characteristics of each of these four clusters revealed three clusters (n=13,133) with samples having biomarker results typical of a PJI-negative classification and one cluster (n=4,032) with samples having biomarker results typical of a PJI-positive classification. A decision tree model, trained and tested independently of external diagnostic rules, was then developed to match the classification determined by the unsupervised clustering. The performance of the model was assessed versus a modified 2018 International Consensus Meeting (ICM) criteria, in both the test cohort and an independent unlabeled validation set of 5,601 samples. The SHAP (SHapley Additive exPlanations) method was used to explore feature importance. Results The ML model showed an area under the curve of 0.993, with a sensitivity of 98.8%, specificity of 97.3%, positive predictive value (PPV) of 92.9%, and negative predictive value (NPV) of 99.8% in predicting the modified 2018 ICM diagnosis among test set samples. The model maintained its diagnostic accuracy in the validation cohort, yielding 99.1% sensitivity, 97.1% specificity, 91.9% PPV, and 99.9% NPV. The model's inconclusive rate (diagnostic probability between 20-80%) in the validation cohort was only 1.3%, lower than that observed with the modified 2018 ICM PJI classification (7.4%; p<0.001). The SHAP analysis found that AD was the most important feature in the model, exhibiting dominance among >95% of "infected" and "not infected" diagnoses. Other important features were the sum of the MID test panel, %PMN, and SF-CRP. Conclusions Although defined methods and tools for diagnosis of PJI using multiple biomarker criteria are available, they are not consistently applied or widely implemented. There is a need for algorithmic interpretation of these biomarkers to enable consistent interpretation of the results to drive treatment decisions. The new model, using clinical parameters measured from a patient's SF sample, renders a preoperative probability score for PJI which performs well compared to a modified 2018 ICM definition. Taken together with other clinical signs, this model has the potential to increase the accuracy of clinical evaluations and reduce the rate of inconclusive classification, thereby enabling more appropriate and expedited downstream treatment decisions.

Clinical Implication of Maumgyeol Basic Service-the 2 Channel Electroencephalography and a Photoplethysmogram-based Mental Health Evaluation Software

Objective

Maumgyeol Basic service is a mental health evaluation and grade scoring software using the 2 channels EEG and photoplethysmogram (PPG). This service is supposed to assess potential at-risk groups with mental illness more easily, rapidly, and reliably. This study aimed to evaluate the clinical implication of the Maumgyeol Basic service.

Methods

One hundred one healthy controls and 103 patients with a psychiatric disorder were recruited. Psychological evaluation (Mental Health Screening for Depressive Disorders [MHS-D], Mental Health Screening for Anxiety Disorders [MHS-A], cognitive stress response scale [CSRS], 12-item General Health Questionnaire [GHQ-12], Clinical Global Impression [CGI]) and digit symbol substitution test (DSST) were applied to all participants. Maumgyeol brain health score and Maumgyeol mind health score were calculated from 2 channel frontal EEG and PPG, respectively.

Results

Participants were divided into three groups: Maumgyeol Risky, Maumgyeol Good, and Maumgyeol Usual. The Maumgyeol mind health scores, but not brain health scores, were significantly lower in the patients group compared to healthy controls. Maumgyeol Risky group showed significantly lower psychological and cognitive ability evaluation scores than Maumgyeol Usual and Good groups. Maumgyel brain health score showed significant correlations with CSRS and DSST. Maumgyeol mind health score showed significant correlations with CGI and DSST. About 20.6% of individuals were classified as the No Insight group, who had mental health problems but were unaware of their illnesses.

Conclusion

This study suggests that the Maumgyeol Basic service can provide important clinical information about mental health and be used as a meaningful digital mental healthcare monitoring solution to prevent symptom aggravation.

Public health nursing workforce and learning needs: A national sample survey analysis

OBJECTIVES

Generate national estimates of the public health nursing workforce's (1) demographic and work characteristics and (2) continuing education learning needs in the United States.

DESIGN

Secondary data analysis of the 2018 National Sample Survey of Registered Nurses.

SAMPLE

Total 7352 of the 50,273 survey respondents were categorized as public health nurses (PHNs), representing an estimated 467,271 national workforce.

MEASUREMENTS

Survey items for demographics, practice setting, training topics, and language(s) spoken fluently were analyzed.

RESULTS

Workforce demographic characteristics are included. Mental health training was the most frequently endorsed topic by PHNs, followed by patient-centered care and evidence-based care. Training topic needs vary by practice setting.

CONCLUSIONS

Results here can be used as a needs assessment for national public health nursing professional development and education initiatives. Further research is needed to refine and survey a nationally representative sample in a manner meaningful to public health nursing practice.

A framework to identify ethical concerns with ML-guided care workflows: a case study of mortality prediction to guide advance care planning

OBJECTIVE

Identifying ethical concerns with ML applications to healthcare (ML-HCA) before problems arise is now a stated goal of ML design oversight groups and regulatory agencies. Lack of accepted standard methodology for ethical analysis, however, presents challenges. In this case study, we evaluate use of a stakeholder "values-collision" approach to identify consequential ethical challenges associated with an ML-HCA for advanced care planning (ACP). Identification of ethical challenges could guide revision and improvement of the ML-HCA.

MATERIALS AND METHODS

We conducted semistructured interviews of the designers, clinician-users, affiliated administrators, and patients, and inductive qualitative analysis of transcribed interviews using modified grounded theory.

RESULTS

Seventeen stakeholders were interviewed. Five "values-collisions"-where stakeholders disagreed about decisions with ethical implications-were identified: (1) end-of-life workflow and how model output is introduced; (2) which stakeholders receive predictions; (3) benefit-harm trade-offs; (4) whether the ML design team has a fiduciary relationship to patients and clinicians; and, (5) how and if to protect early deployment research from external pressures, like news scrutiny, before research is completed.

DISCUSSION

From these findings, the ML design team prioritized: (1) alternative workflow implementation strategies; (2) clarification that prediction was only evaluated for ACP need, not other mortality-related ends; and (3) shielding research from scrutiny until endpoint driven studies were completed.

CONCLUSION

In this case study, our ethical analysis of this ML-HCA for ACP was able to identify multiple sites of intrastakeholder disagreement that mark areas of ethical and value tension. These findings provided a useful initial ethical screening.