Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Information Has Been Removed With Use of Machine Learning

OptimCLM: Optimizing clinical language models for predicting patient outcomes via knowledge distillation, pruning and quantization

BACKGROUND

Clinical Language Models (CLMs) possess the potential to reform traditional healthcare systems by aiding in clinical decision making and optimal resource utilization. They can enhance patient outcomes and help healthcare management through predictive clinical tasks. However, their real-world deployment is limited due to high computational cost at inference, in terms of both time and space complexity.

OBJECTIVE

This study aims to develop and optimize an efficient framework that compresses CLMs without significant performance loss, reducing inference time and disk-space, and enabling real-world clinical applications.

METHODS

We introduce OptimCLM, a framework for optimizing CLMs with ensemble learning, knowledge distillation (KD), pruning and quantization. Based on domain-knowledge and performance, we select and combine domain-adaptive CLMs DischargeBERT and COReBERT as the teacher ensemble model. We transfer the teacher's knowledge to two smaller generalist models, BERT-PKD and TinyBERT, and apply black-box KD, post-training unstructured pruning and post-training 8-bit model quantization to them. In an admission-to-discharge setting, we evaluate the framework on four clinical outcome prediction tasks (length of stay prediction, mortality prediction, diagnosis prediction and procedure prediction) using admission notes from the MIMIC-III clinical database.

RESULTS

The OptimCLM framework achieved up to 22.88× compression ratio and 28.7× inference speedup, with less than 5% and 2% loss in macro-averaged AUROC for TinyBERT and BERT-PKD, respectively. The teacher model outperformed five state-of-the-art models on all tasks. The optimized BERT-PKD model also outperformed them in most tasks.

CONCLUSION

Our findings suggest that domain-specific fine-tuning with ensemble learning and KD is more effective than domain-specific pre-training for domain-knowledge transfer and text classification tasks. Thus, this work demonstrates the feasibility and potential of deploying optimized CLMs in healthcare settings and developing them with less computational resources.

The potential for large language models to transform cardiovascular medicine

Cardiovascular diseases persist as the leading cause of death globally and their early detection and prediction remain a major challenge. Artificial intelligence (AI) tools can help meet this challenge as they have considerable potential for early diagnosis and prediction of occurrence of these diseases. Deep neural networks can improve the accuracy of medical image interpretation and their outputs can provide rich information that otherwise would not be detected by cardiologists. With recent advances in transformer models, multimodal AI, and large language models, the ability to integrate electronic health record data with images, genomics, biosensors, and other data has the potential to improve diagnosis and partition patients who are at high risk for primary preventive strategies. Although much emphasis has been placed on AI supporting clinicians, AI can also serve patients and provide immediate help with diagnosis, such as that of arrhythmia, and is being studied for automated self-imaging. Potential risks, such as loss of data privacy or potential diagnostic errors, should be addressed before use in clinical practice. This Series paper explores opportunities and limitations of AI models for cardiovascular medicine, and aims to identify specific barriers to and solutions in the application of AI models, facilitating their integration into health-care systems.

Multimodal Large Language Models in Health Care: Applications, Challenges, and Future Outlook

In the complex and multidimensional field of medicine, multimodal data are prevalent and crucial for informed clinical decisions. Multimodal data span a broad spectrum of data types, including medical images (eg, MRI and CT scans), time-series data (eg, sensor data from wearable devices and electronic health records), audio recordings (eg, heart and respiratory sounds and patient interviews), text (eg, clinical notes and research articles), videos (eg, surgical procedures), and omics data (eg, genomics and proteomics). While advancements in large language models (LLMs) have enabled new applications for knowledge retrieval and processing in the medical field, most LLMs remain limited to processing unimodal data, typically text-based content, and often overlook the importance of integrating the diverse data modalities encountered in clinical practice. This paper aims to present a detailed, practical, and solution-oriented perspective on the use of multimodal LLMs (M-LLMs) in the medical field. Our investigation spanned M-LLM foundational principles, current and potential applications, technical and ethical challenges, and future research directions. By connecting these elements, we aimed to provide a comprehensive framework that links diverse aspects of M-LLMs, offering a unified vision for their future in health care. This approach aims to guide both future research and practical implementations of M-LLMs in health care, positioning them as a paradigm shift toward integrated, multimodal data-driven medical practice. We anticipate that this work will spark further discussion and inspire the development of innovative approaches in the next generation of medical M-LLM systems.

Predictive modeling of lean body mass, appendicular lean mass, and appendicular skeletal muscle mass using machine learning techniques: A comprehensive analysis utilizing NHANES data and the Look AHEAD study

This study addresses the pressing need for improved methods to predict lean mass in adults, and in particular lean body mass (LBM), appendicular lean mass (ALM), and appendicular skeletal muscle mass (ASMM) for the early detection and management of sarcopenia, a condition characterized by muscle loss and dysfunction. Sarcopenia presents significant health risks, especially in populations with chronic diseases like cancer and the elderly. Current assessment methods, primarily relying on Dual-energy X-ray absorptiometry (DXA) scans, lack widespread applicability, hindering timely intervention. Leveraging machine learning techniques, this research aimed to develop and validate predictive models using data from the National Health and Nutrition Examination Survey (NHANES) and the Action for Health in Diabetes (Look AHEAD) study. The models were trained on anthropometric data, demographic factors, and DXA-derived metrics to accurately estimate LBM, ALM, and ASMM normalized to weight. Results demonstrated consistent performance across various machine learning algorithms, with LassoNet, a non-linear extension of the popular LASSO method, exhibiting superior predictive accuracy. Notably, the integration of bone mineral density measurements into the models had minimal impact on predictive accuracy, suggesting potential alternatives to DXA scans for lean mass assessment in the general population. Despite the robustness of the models, limitations include the absence of outcome measures and cohorts highly vulnerable to muscle mass loss. Nonetheless, these findings hold promise for revolutionizing lean mass assessment paradigms, offering implications for chronic disease management and personalized health interventions. Future research endeavors should focus on validating these models in diverse populations and addressing clinical complexities to enhance prediction accuracy and clinical utility in managing sarcopenia.

Ethical Implications of Artificial Intelligence in Gastroenterology: The Co-pilot or the Captain?

Though artificial intelligence (AI) is being widely implemented in gastroenterology (GI) and hepatology and has the potential to be paradigm shifting for clinical practice, its pitfalls must be considered along with its advantages. Currently, although the use of AI is limited in practice to supporting clinical judgment, medicine is rapidly heading toward a global environment where AI will be increasingly autonomous. Broader implementation of AI will require careful ethical considerations, specifically related to bias, privacy, and consent. Widespread use of AI raises concerns related to increasing rates of systematic errors, potentially due to bias introduced in training datasets. We propose that a central repository for collection and analysis for training and validation datasets is essential to overcoming potential biases. Since AI does not have built-in concepts of bias and equality, humans involved in AI development and implementation must ensure its ethical use and development. Moreover, ethical concerns regarding data ownership and health information privacy are likely to emerge, obviating traditional methods of obtaining patient consent that cover all possible uses of patient data. The question of liability in case of adverse events related to use of AI in GI must be addressed among the physician, the healthcare institution, and the AI developer. Though the future of AI in GI is very promising, herein we review the ethical considerations in need of additional guidance informed by community experience and collective expertise.

Transforming Cardiovascular Care With Artificial Intelligence: From Discovery to Practice: JACC State-of-the-Art Review

Artificial intelligence (AI) has the potential to transform every facet of cardiovascular practice and research. The exponential rise in technology powered by AI is defining new frontiers in cardiovascular care, with innovations that span novel diagnostic modalities, new digital native biomarkers of disease, and high-performing tools evaluating care quality and prognosticating clinical outcomes. These digital innovations promise expanded access to cardiovascular screening and monitoring, especially among those without access to high-quality, specialized care historically. Moreover, AI is propelling biological and clinical discoveries that will make future cardiovascular care more personalized, precise, and effective. The review brings together these diverse AI innovations, highlighting developments in multimodal cardiovascular AI across clinical practice and biomedical discovery, and envisioning this new future backed by contemporary science and emerging discoveries. Finally, we define the critical path and the safeguards essential to realizing this AI-enabled future that helps achieve optimal cardiovascular health and outcomes for all.

HydraGAN: A Cooperative Agent Model for Multi-Objective Data Generation

Generative adversarial networks have become a de facto approach to generate synthetic data points that resemble their real counterparts. We tackle the situation where the realism of individual samples is not the sole criterion for synthetic data generation. Additional constraints such as privacy preservation, distribution realism, and diversity promotion may also be essential to optimize. To address this challenge, we introduce HydraGAN, a multi-agent network that performs multi-objective synthetic data generation. We theoretically verify that training the HydraGAN system, containing a single generator and an arbitrary number of discriminators, leads to a Nash equilibrium. Experimental results for six datasets indicate that HydraGAN consistently outperforms prior methods in maximizing the Area under the Radar Curve (AuRC), balancing a combination of cooperative or competitive data generation goals.

User experience of a family health history chatbot: A quantitative analysis

OBJECTIVE

Family health history (FHx) is an important tool in assessing one's risk towards specific health conditions. However, user experience of FHx collection tools is rarely studied. ItRunsInMyFamily.com (ItRuns) was developed to assess FHx and hereditary cancer risk. This study reports a quantitative user experience analysis of ItRuns.

METHODS

We conducted a public health campaign in November 2019 to promote FHx collection using ItRuns. We used software telemetry to quantify abandonment and time spent on ItRuns to identify user behaviors and potential areas of improvement.

RESULTS

Of 11,065 users who started the ItRuns assessment, 4305 (38.91%) reached the final step to receive recommendations about hereditary cancer risk. Highest abandonment rates were during Introduction (32.82%), Invite Friends (29.03%), and Family Cancer History (12.03%) subflows. Median time to complete the assessment was 636 s. Users spent the highest median time on Proband Cancer History (124.00 s) and Family Cancer History (119.00 s) subflows. Search list questions took the longest to complete (median 19.50 s), followed by free text email input (15.00 s).

CONCLUSION

Knowledge of objective user behaviors at a large scale and factors impacting optimal user experience will help enhance the ItRuns workflow and improve future FHx collection.

The Utilization of e-Health in Lymphedema Care: A Narrative Review

Background: Electronic health (e-Health), refers to technologies that can be utilized to enhance patient care as well as collect and share health information. e-Health comprises several umbrella terms, including telehealth, mobile health, e-Health, wearables, and artificial intelligence. The types of e-Health technologies being utilized in lymphedema (LE) care are unknown. Method: In this narrative review, a search of published research on the utilization of e-Health technologies in LE-related care was conducted. Results: Five different types of e-Health modalities were found (robotics, artificial intelligence, electronic medical records, smart wearable devices, and instructive online information) spanning 14 use cases and 4 phases of care (preventative, diagnostic, assessment, and treatment phases). Broad e-Health utilization examples were found including robotic-assisted surgery to reduce the likelihood of LE after lymphadenectomy, machine learning to predict patients at risk of filarial-related LE, and a novel wearable device prototype designed to provide lymphatic drainage. Conclusions: e-Health has reported merit in the prevention, diagnoses, assessment, and treatment of LE with utilization demonstrating cutting edge applicability of e-Health for achieving optimal patient care and outcomes. As technology continues to advance, additional research into the utilization of e-Health in LE care is warranted.

Navigating ethical challenges in psychological research involving digital remote technologies and people who use alcohol or drugs

Digital and remote technologies (DRT) are increasingly being used in scientific investigations to objectively measure human behavior during day-to-day activities. Using these devices, psychologists and other behavioral scientists can investigate health risk behaviors, such as drug and alcohol use, by closely examining the causes and consequences of monitored behaviors as they occur naturalistically. There are, however, complex ethical issues that emerge when using DRT methodologies in research with people who use substances. These issues must be identified and addressed so DRT devices can be incorporated into psychological research with this population in a manner that comports the ethical standards of the American Psychological Association. In this article, we discuss the ethical ramifications of using DRT in behavioral studies with people who use substances. Drawing on allied fields with similar ethical issues, we make recommendations to researchers who wish to incorporate DRT into their own research. Major topics include (a) threats to and methods for protecting participant and nonparticipant privacy, (b) shortcomings of traditional informed consent in DRT research, (c) researcher liabilities introduced by real-time continuous data collection, (d) threats to distributive justice arising from computational tools often used to manage and analyze DRT data, and (e) ethical implications of the "digital divide." We conclude with a more optimistic discussion of how DRT may provide safer alternatives to gold standard paradigms in substance use research, allowing researchers to test hypotheses that were previously prohibited on ethical grounds. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Artificial intelligence (AI) or augmented intelligence? How big data and AI are transforming healthcare: Challenges and opportunities

The sanctity of the doctor-patient relationship is deeply embedded in tradition - the Hippocratic oath, medical ethics, professional codes of conduct, and legislation - all of which are being disrupted by big data and 'artificial' intelligence (AI). The transition from paper-based records to electronic health records, wearables, mobile health applications and mobile phone data has created new opportunities to scale up data collection. Databases of unimaginable magnitude can be harnessed to develop algorithms for AI and to refine machine learning. Complex neural networks now lie at the core of ubiquitous AI systems in healthcare. A transformed healthcare environment enhanced by innovation, robotics, digital technology, and improved diagnostics and therapeutics is plagued by ethical, legal and social challenges. Global guidelines are emerging to ensure governance in AI, but many low- and middle-income countries have yet to develop context- specific frameworks. Legislation must be developed to frame liability and account for negligence due to robotics in the same way human healthcare providers are held accountable. The digital divide between high- and low-income settings is significant and has the potential to exacerbate health inequities globally.

Deep Learning for Epidemiologists: An Introduction to Neural Networks

Deep learning methods are increasingly being applied to problems in medicine and health care. However, few epidemiologists have received formal training in these methods. To bridge this gap, this article introduces the fundamentals of deep learning from an epidemiologic perspective. Specifically, this article reviews core concepts in machine learning (e.g., overfitting, regularization, and hyperparameters); explains several fundamental deep learning architectures (convolutional neural networks, recurrent neural networks); and summarizes training, evaluation, and deployment of models. Conceptual understanding of supervised learning algorithms is the focus of the article; instructions on the training of deep learning models and applications of deep learning to causal learning are out of this article's scope. We aim to provide an accessible first step towards enabling the reader to read and assess research on the medical applications of deep learning and to familiarize readers with deep learning terminology and concepts to facilitate communication with computer scientists and machine learning engineers.

Data Privacy in Healthcare: In the Era of Artificial Intelligence

Data Privacy has increasingly become a matter of concern in the era of large public digital respositories of data. This is particularly true in healthcare where data can be misused if traced back to patients, and brings with itself a myriad of possibilities. Bring custodians of data, as well as being at the helm of disigning studies and products that can potentially benefit products, healthcare professionals often find themselves unsure about ethical and legal constraints that undelie data sharing. In this review we touch upon the concerns, leal frameworks as well as some common practices in these respects.

Innovations in Technology and Mechanisms of Change in Behavioral Interventions

The purpose of this manuscript is to provide an overview of, and rationale for, the increasing adoption of a wide range of cutting-edge technological methods in assessment and intervention which are relevant for treatment. First, we review traditional approaches to measuring and monitoring affect, behavior, and cognition in behavior and cognitive-behavioral therapy. Second, we describe evolving active and passive technology-enabled approaches to behavior assessment including emerging applications of digital phenotyping facilitated through fitness trackers, smartwatches, and social media. Third, we describe ways that these emerging technologies may be used for intervention, focusing on novel applications for the use of technology in intervention efforts. Importantly, though some of the methods and approaches we describe here warrant future testing, many aspects of technology can already be easily incorporated within an established treatment framework.

Artificial intelligence-based risk stratification, accurate diagnosis and treatment prediction in gynecologic oncology

As data-driven science, artificial intelligence (AI) has paved a promising path toward an evolving health system teeming with thrilling opportunities for precision oncology. Notwithstanding the tremendous success of oncological AI in such fields as lung carcinoma, breast tumor and brain malignancy, less attention has been devoted to investigating the influence of AI on gynecologic oncology. Hereby, this review sheds light on the ever-increasing contribution of state-of-the-art AI techniques to the refined risk stratification and whole-course management of patients with gynecologic tumors, in particular, cervical, ovarian and endometrial cancer, centering on information and features extracted from clinical data (electronic health records), cancer imaging including radiological imaging, colposcopic images, cytological and histopathological digital images, and molecular profiling (genomics, transcriptomics, metabolomics and so forth). However, there are still noteworthy challenges beyond performance validation. Thus, this work further describes the limitations and challenges faced in the real-word implementation of AI models, as well as potential solutions to address these issues.

Introduction and Comparison of Novel Decentral Learning Schemes with Multiple Data Pools for Privacy-Preserving ECG Classification

Artificial intelligence and machine learning have led to prominent and spectacular innovations in various scenarios. Application in medicine, however, can be challenging due to privacy concerns and strict legal regulations. Methods that centralize knowledge instead of data could address this issue. In this work, 6 different decentralized machine learning algorithms are applied to 12-lead ECG classification and compared to conventional, centralized machine learning. The results show that state-of-the-art federated learning leads to reasonable losses of classification performance compared to a standard, central model (-0.054 AUROC) while providing a significantly higher level of privacy. A proposed weighted variant of federated learning (-0.049 AUROC) and an ensemble (-0.035 AUROC) outperformed the standard federated learning algorithm. Overall, considering multiple metrics, the novel batch-wise sequential learning scheme performed best (-0.036 AUROC to baseline). Although, the technical aspects of implementing them in a real-world application are to be carefully considered, the described algorithms constitute a way forward towards preserving-preserving AI in medicine.

Large language models in health care: Development, applications, and challenges

Recently, the emergence of ChatGPT, an artificial intelligence chatbot developed by OpenAI, has attracted significant attention due to its exceptional language comprehension and content generation capabilities, highlighting the immense potential of large language models (LLMs). LLMs have become a burgeoning hotspot across many fields, including health care. Within health care, LLMs may be classified into LLMs for the biomedical domain and LLMs for the clinical domain based on the corpora used for pre-training. In the last 3 years, these domain-specific LLMs have demonstrated exceptional performance on multiple natural language processing tasks, surpassing the performance of general LLMs as well. This not only emphasizes the significance of developing dedicated LLMs for the specific domains, but also raises expectations for their applications in health care. We believe that LLMs may be used widely in preconsultation, diagnosis, and management, with appropriate development and supervision. Additionally, LLMs hold tremendous promise in assisting with medical education, medical writing and other related applications. Likewise, health care systems must recognize and address the challenges posed by LLMs.

Systematic review and meta-analysis of the effectiveness of chatbots on lifestyle behaviours

Chatbots (also known as conversational agents and virtual assistants) offer the potential to deliver healthcare in an efficient, appealing and personalised manner. The purpose of this systematic review and meta-analysis was to evaluate the efficacy of chatbot interventions designed to improve physical activity, diet and sleep. Electronic databases were searched for randomised and non-randomised controlled trials, and pre-post trials that evaluated chatbot interventions targeting physical activity, diet and/or sleep, published before 1 September 2022. Outcomes were total physical activity, steps, moderate-to-vigorous physical activity (MVPA), fruit and vegetable consumption, sleep quality and sleep duration. Standardised mean differences (SMD) were calculated to compare intervention effects. Subgroup analyses were conducted to assess chatbot type, intervention type, duration, output and use of artificial intelligence. Risk of bias was assessed using the Effective Public Health Practice Project Quality Assessment tool. Nineteen trials were included. Sample sizes ranged between 25-958, and mean participant age ranged between 9-71 years. Most interventions (n = 15, 79%) targeted physical activity, and most trials had a low-quality rating (n = 14, 74%). Meta-analysis results showed significant effects (all p < 0.05) of chatbots for increasing total physical activity (SMD = 0.28 [95% CI = 0.16, 0.40]), daily steps (SMD = 0.28 [95% CI = 0.17, 0.39]), MVPA (SMD = 0.53 [95% CI = 0.24, 0.83]), fruit and vegetable consumption (SMD = 0.59 [95% CI = 0.25, 0.93]), sleep duration (SMD = 0.44 [95% CI = 0.32, 0.55]) and sleep quality (SMD = 0.50 [95% CI = 0.09, 0.90]). Subgroup analyses showed that text-based, and artificial intelligence chatbots were more efficacious than speech/voice chatbots for fruit and vegetable consumption, and multicomponent interventions were more efficacious than chatbot-only interventions for sleep duration and sleep quality (all p < 0.05). Findings from this systematic review and meta-analysis indicate that chatbot interventions are efficacious for increasing physical activity, fruit and vegetable consumption, sleep duration and sleep quality. Chatbot interventions were efficacious across a range of populations and age groups, with both short- and longer-term interventions, and chatbot only and multicomponent interventions being efficacious.

Fingerprinting walking using wrist-worn accelerometers

BACKGROUND

Identifying an individual from accelerometry data collected during walking without reliance on step-cycle detection has not been achieved with high accuracy.

RESEARCH QUESTION

We propose an open-source reproducible method to: (1) create a unique, person-specific "walking fingerprint" from a sample of un-landmarked high-resolution data collected by a wrist-worn accelerometer; and (2) predict who an individual is from their walking fingerprint.

METHODS

Accelerometry data were collected during walking from 32 individuals (23-52 y.o., 19 females) for at least 380 s each. For this study's purpose, data are not landmarked, nor synchronized. Individual walking fingerprints were created by: (1) partitioning the accelerometer time series in adjacent, non-overlapping one-second intervals; (2) transforming all one-second interval data for a given individual into a three-dimensional (3D) image obtained by plotting each one-second interval time series by the lagged time series for a series of lags; (3) partitioning these resulting participant-specific 3D images into a grid of cells; and (4) identifying the combinations of cells (areas in the 3D image) that best predict the individual. For every participant, the first 200 s of data were used as training and the last 180 s as testing. This approach does not use segmentation methods for individual strides, which reduces dependence on complementary algorithms and increases its generalizability.

RESULTS

The method correctly identified 100 % of the participants in the test data and highlighted unique features of walking that characterize the individuals.

SIGNIFICANCE

Predicting the identity of an individual from their walking pattern has immediate implications that can complement or replace those of actual fingerprinting, voice, and image recognition. Furthermore, as walking may change with age or disease burden, individual walking fingerprints may be used as biomarkers of change in health status with potential clinical and epidemiologic implications.

User experience of a family health history chatbot: A quantitative analysis

Objective

Family health history (FHx) is an important tool in assessing one's risk towards specific health conditions. However, user experience of FHx collection tools is rarely studied. ItRunsInMyFamily.com (ItRuns) was developed to assess FHx and hereditary cancer risk. This study reports a quantitative user experience analysis of ItRuns.

Methods

We conducted a public health campaign in November 2019 to promote FHx collection using ItRuns. We used software telemetry to quantify abandonment and time spent on ItRuns to identify user behaviors and potential areas of improvement.

Results

Of 11065 users who started the ItRuns assessment, 4305 (38.91%) reached the final step to receive recommendations about hereditary cancer risk. Highest abandonment rates were during Introduction (32.82%), Invite Friends (29.03%), and Family Cancer History (12.03%) subflows. Median time to complete the assessment was 636 seconds. Users spent the highest median time on Proband Cancer History (124.00 seconds) and Family Cancer History (119.00 seconds) subflows. Search list questions took the longest to complete (median 19.50 seconds), followed by free text email input (15.00 seconds).

Conclusion

Knowledge of objective user behaviors at a large scale and factors impacting optimal user experience will help enhance the ItRuns workflow and improve future FHx collection.

Should artificial intelligence be used in conjunction with Neuroimaging in the diagnosis of Alzheimer's disease?

Alzheimer's disease (AD) is a progressive, neurodegenerative disorder that affects memory, thinking, behavior, and other cognitive functions. Although there is no cure, detecting AD early is important for the development of a therapeutic plan and a care plan that may preserve cognitive function and prevent irreversible damage. Neuroimaging, such as magnetic resonance imaging (MRI), computed tomography (CT), and positron emission tomography (PET), has served as a critical tool in establishing diagnostic indicators of AD during the preclinical stage. However, as neuroimaging technology quickly advances, there is a challenge in analyzing and interpreting vast amounts of brain imaging data. Given these limitations, there is great interest in using artificial Intelligence (AI) to assist in this process. AI introduces limitless possibilities in the future diagnosis of AD, yet there is still resistance from the healthcare community to incorporate AI in the clinical setting. The goal of this review is to answer the question of whether AI should be used in conjunction with neuroimaging in the diagnosis of AD. To answer the question, the possible benefits and disadvantages of AI are discussed. The main advantages of AI are its potential to improve diagnostic accuracy, improve the efficiency in analyzing radiographic data, reduce physician burnout, and advance precision medicine. The disadvantages include generalization and data shortage, lack of in vivo gold standard, skepticism in the medical community, potential for physician bias, and concerns over patient information, privacy, and safety. Although the challenges present fundamental concerns and must be addressed when the time comes, it would be unethical not to use AI if it can improve patient health and outcome.

Challenges to Using Big Data in Cancer

Big data in healthcare can enable unprecedented understanding of diseases and their treatment, particularly in oncology. These data may include electronic health records, medical imaging, genomic sequencing, payor records, and data from pharmaceutical research, wearables, and medical devices. The ability to combine datasets and use data across many analyses is critical to the successful use of big data and is a concern for those who generate and use the data. Interoperability and data quality continue to be major challenges when working with different healthcare datasets. Mapping terminology across datasets, missing and incorrect data, and varying data structures make combining data an onerous and largely manual undertaking. Data privacy is another concern addressed by the Health Insurance Portability and Accountability Act, the Common Rule, and the General Data Protection Regulation. The use of big data is now included in the planning and activities of the FDA and the European Medicines Agency. The willingness of organizations to share data in a precompetitive fashion, agreements on data quality standards, and institution of universal and practical tenets on data privacy will be crucial to fully realizing the potential for big data in medicine.

Leveraging Mobile Technology for Public Health Promotion: A Multidisciplinary Perspective

Health behaviors are inextricably linked to health and well-being, yet issues such as physical inactivity and insufficient sleep remain significant global public health problems. Mobile technology-and the unprecedented scope and quantity of data it generates-has a promising but largely untapped potential to promote health behaviors at the individual and population levels. This perspective article provides multidisciplinary recommendations on the design and use of mobile technology, and the concomitant wealth of data, to promote behaviors that support overall health. Using physical activity as anexemplar health behavior, we review emerging strategies for health behavior change interventions. We describe progress on personalizing interventions to an individual and their social, cultural, and built environments, as well as on evaluating relationships between mobile technology data and health to establish evidence-based guidelines. In reviewing these strategies and highlighting directions for future research, we advance the use of theory-based, personalized, and human-centered approaches in promoting health behaviors.

Reconsenting paediatric research participants for use of identifying data

When a minor research participant reaches the age of majority or the level of maturity necessary to be granted legal decision-making capacity, reconsent can be required for ongoing participation in research or use of health information and banked biological materials. Despite potential logistical concerns with implementation and ethical questions about the trade-offs between maximising respect for participant agency and facilitating research that may generate benefits, reconsent is the approach most consistent with both law and research ethics.Canadian common law consent requirements are expansive and likely compel reconsent on obtaining capacity. Common law doctrine recognises that children are entitled to decision-making authority that reflects their evolving intelligence and understanding. Health consent legislation varies by province but generally either compels reconsent on obtaining capacity or delegates the ability to determine reconsent to research ethics boards. These boards largely rely on the Canada's national ethics policy, the Tri-Council Policy Statement, which states that, with few exceptions, reconsent for continued participation is required when minors gain capacity that would allow them to consent to the research in which they participate. A strict interpretation of this policy could require researchers to perform frequent capacity assessments, potentially presenting feasibility concerns. In addition, Canadian policy and law are generally consistent with the core principles of key international ethical standards from the United Nations and elsewhere.In sum, reconsent of paediatric participants upon obtaining capacity should be explicit and informed in Canada, and should not be presumed from continued participation alone.

Synthetic data in health care: A narrative review

Data are central to research, public health, and in developing health information technology (IT) systems. Nevertheless, access to most data in health care is tightly controlled, which may limit innovation, development, and efficient implementation of new research, products, services, or systems. Using synthetic data is one of the many innovative ways that can allow organizations to share datasets with broader users. However, only a limited set of literature is available that explores its potentials and applications in health care. In this review paper, we examined existing literature to bridge the gap and highlight the utility of synthetic data in health care. We searched PubMed, Scopus, and Google Scholar to identify peer-reviewed articles, conference papers, reports, and thesis/dissertations articles related to the generation and use of synthetic datasets in health care. The review identified seven use cases of synthetic data in health care: a) simulation and prediction research, b) hypothesis, methods, and algorithm testing, c) epidemiology/public health research, d) health IT development, e) education and training, f) public release of datasets, and g) linking data. The review also identified readily and publicly accessible health care datasets, databases, and sandboxes containing synthetic data with varying degrees of utility for research, education, and software development. The review provided evidence that synthetic data are helpful in different aspects of health care and research. While the original real data remains the preferred choice, synthetic data hold possibilities in bridging data access gaps in research and evidence-based policymaking.

Progress towards unique patient identification and case-based surveillance within the Southern African development community

Population mobility makes patient-tracking and care linkage in the South African Development Community (SADC) challenging. Case-based surveillance (CBS) through individual-level clinical data linked with a unique patient-identifier (UPI) is recommended. We conducted a mixed-methods landscape analysis of UPI and CBS implementation within selected SADC countries, this included: (1) SADC UPI implementation literature review; (2) assessment of UPI and CBS implementation for high HIV-prevalence SADC countries; (3) UPI implementation case-study in selected South African primary healthcare (PHC) facilities. Research into CBS and UPI implementation for the SADC region is lacking. Existing patient-identification methods often fail and limit patient-tracking. Paper-based records and poor integration between health-information systems further restrict patient-tracking. Most countries were in the early-middle stages of CBS and faced UPI challenges. Our South African case-study found that the UPI often goes uncaptured. Difficulties tracking patients across prevention and care cascades will continue until a functional and reliable UPI is available.

Respect, justice and learning are limited when patients are deidentified data subjects

Introduction

Critical for advancing a Learning Health System (LHS) in the U.S., a regulatory safe harbor for deidentified data reduces barriers to learning from care at scale while minimizing privacy risks. We examine deidentified data policy as a mechanism for synthesizing the ethical obligations underlying clinical care and human subjects research for an LHS which conceptually and practically integrates care and research, blurring the roles of patient and subject.

Methods

First, we discuss respect for persons vis-a-vis the systemic secondary use of data and tissue collected in the fiduciary context of clinical care. We argue that, without traditional informed consent or duty to benefit the individual, deidentification may allow secondary use to supersede the primary purpose of care. Next, we consider the effectiveness of deidentification for minimizing harms via privacy protection and maximizing benefits via promoting learning and translational care. We find that deidentification is unable to fully protect privacy given the vastness of health data and current technology, yet it imposes limitations to learning and barriers for efficient translation. After that, we evaluate the impact of deidentification on distributive justice within an LHS ethical framework in which patients are obligated to contribute to learning and the system has a duty to translate knowledge into better care. Such a system may permit exacerbation of health disparities as it accelerates learning without mechanisms to ensure that individuals' contributions and benefits are fair and balanced.

Results

We find that, despite its established advantages, system-wide use of deidentification may be suboptimal for signaling respect, protecting privacy or promoting learning, and satisfying requirements of justice for patients and subjects.

Conclusions

Finally, we highlight ethical, socioeconomic, technological and legal challenges and next steps, including a critical appreciation for novel approaches to realize an LHS that maximizes efficient, effective learning and just translation without the compromises of deidentification.

Big Health Data Research and Group Harm: The Scope of IRB Review

Much of precision medicine is driven by big health data research-the analysis of massive datasets representing the complex web of genetic, behavioral, environmental, and other factors that impact human well-being. There are some who point to the Common Rule, the regulation governing federally funded human subjects research, as a regulatory panacea for all types of big health data research. But how well does the Common Rule fit the regulatory needs of this type of research? This article suggests that harms that may arise from artificial intelligence and machine-learning technologies used in big health data research-and the increased likelihood that this research will affect public policy-mean it is time to consider whether the current human research regulations prohibit comprehensive, ethical review of big health data research that may result in group harm.

Selecting Privacy-Enhancing Technologies for Managing Health Data Use

Privacy protection for health data is more than simply stripping datasets of specific identifiers. Privacy protection increasingly means the application of privacy-enhancing technologies (PETs), also known as privacy engineering. Demands for the application of PETs are not yet met with ease of use or even understanding. This paper provides a scope of the current peer-reviewed evidence regarding the practical use or adoption of various PETs for managing health data privacy. We describe the state of knowledge of PETS for the use and exchange of health data specifically and build a practical perspective on the steps needed to improve the standardization of the application of PETs for diverse uses of health data.

Passive Data Use for Ethical Digital Public Health Surveillance in a Postpandemic World

There is a fundamental need to establish the most ethical and effective way of tracking disease in the postpandemic era. The ubiquity of mobile phones is generating large amounts of passive data (collected without active user participation) that can be used as a tool for tracking disease. Although discussions of pragmatism or economic issues tend to guide public health decisions, ethical issues are the foremost public concern. Thus, officials must look to history and current moral frameworks to avoid past mistakes and ethical pitfalls. Past pandemics demonstrate that the aftermath is the most effective time to make health policy decisions. However, an ethical discussion of passive data use for digital public health surveillance has yet to be attempted, and little has been done to determine the best method to do so. Therefore, we aim to highlight four potential areas of ethical opportunity and challenge: (1) informed consent, (2) privacy, (3) equity, and (4) ownership.

Nonfungible Tokens as a Blockchain Solution to Ethical Challenges for the Secondary Use of Biospecimens: Viewpoint

Henrietta Lacks' deidentified tissue became HeLa cells (the paradigmatic learning health platform). In this article, we discuss separating research on Ms Lacks' tissue from obligations to promote respect, beneficence, and justice for her as a patient. This case illuminates ethical challenges for the secondary use of biospecimens, which persist in contemporary learning health systems. Deidentification and broad consent seek to maximize the benefits of learning from care by minimizing burdens on patients, but these strategies are insufficient for privacy, transparency, and engagement. The resulting supply chain for human cellular and tissue-based products may therefore recapitulate the harms experienced by the Lacks family. We introduce the potential for blockchain technology to build unprecedented transparency, engagement, and accountability into learning health system architecture without requiring deidentification. The ability of nonfungible tokens to maintain the provenance of inherently unique digital assets may optimize utility, value, and respect for patients who contribute tissue and other clinical data for research. We consider the potential benefits and survey major technical, ethical, socioeconomic, and legal challenges for the successful implementation of the proposed solutions. The potential for nonfungible tokens to promote efficiency, effectiveness, and justice in learning health systems demands further exploration.

Human, All Too Human? An All-Around Appraisal of the "Artificial Intelligence Revolution" in Medical Imaging

Artificial intelligence (AI) has seen dramatic growth over the past decade, evolving from a niche super specialty computer application into a powerful tool which has revolutionized many areas of our professional and daily lives, and the potential of which seems to be still largely untapped. The field of medicine and medical imaging, as one of its various specialties, has gained considerable benefit from AI, including improved diagnostic accuracy and the possibility of predicting individual patient outcomes and options of more personalized treatment. It should be noted that this process can actively support the ongoing development of advanced, highly specific treatment strategies (e.g., target therapies for cancer patients) while enabling faster workflow and more efficient use of healthcare resources. The potential advantages of AI over conventional methods have made it attractive for physicians and other healthcare stakeholders, raising much interest in both the research and the industry communities. However, the fast development of AI has unveiled its potential for disrupting the work of healthcare professionals, spawning concerns among radiologists that, in the future, AI may outperform them, thus damaging their reputations or putting their jobs at risk. Furthermore, this development has raised relevant psychological, ethical, and medico-legal issues which need to be addressed for AI to be considered fully capable of patient management. The aim of this review is to provide a brief, hopefully exhaustive, overview of the state of the art of AI systems regarding medical imaging, with a special focus on how AI and the entire healthcare environment should be prepared to accomplish the goal of a more advanced human-centered world.

Overcoming barriers to data sharing with medical image generation: a comprehensive evaluation

Privacy concerns around sharing personally identifiable information are a major barrier to data sharing in medical research. In many cases, researchers have no interest in a particular individual's information but rather aim to derive insights at the level of cohorts. Here, we utilise generative adversarial networks (GANs) to create medical imaging datasets consisting entirely of synthetic patient data. The synthetic images ideally have, in aggregate, similar statistical properties to those of a source dataset but do not contain sensitive personal information. We assess the quality of synthetic data generated by two GAN models for chest radiographs with 14 radiology findings and brain computed tomography (CT) scans with six types of intracranial haemorrhages. We measure the synthetic image quality by the performance difference of predictive models trained on either the synthetic or the real dataset. We find that synthetic data performance disproportionately benefits from a reduced number of classes. Our benchmark also indicates that at low numbers of samples per class, label overfitting effects start to dominate GAN training. We conducted a reader study in which trained radiologists discriminate between synthetic and real images. In accordance with our benchmark results, the classification accuracy of radiologists improves with an increasing resolution. Our study offers valuable guidelines and outlines practical conditions under which insights derived from synthetic images are similar to those that would have been derived from real data. Our results indicate that synthetic data sharing may be an attractive alternative to sharing real patient-level data in the right setting.

Privacy and artificial intelligence: challenges for protecting health information in a new era

Background

Advances in healthcare artificial intelligence (AI) are occurring rapidly and there is a growing discussion about managing its development. Many AI technologies end up owned and controlled by private entities. The nature of the implementation of AI could mean such corporations, clinics and public bodies will have a greater than typical role in obtaining, utilizing and protecting patient health information. This raises privacy issues relating to implementation and data security.

Main body

The first set of concerns includes access, use and control of patient data in private hands. Some recent public–private partnerships for implementing AI have resulted in poor protection of privacy. As such, there have been calls for greater systemic oversight of big data health research. Appropriate safeguards must be in place to maintain privacy and patient agency. Private custodians of data can be impacted by competing goals and should be structurally encouraged to ensure data protection and to deter alternative use thereof. Another set of concerns relates to the external risk of privacy breaches through AI-driven methods. The ability to deidentify or anonymize patient health data may be compromised or even nullified in light of new algorithms that have successfully reidentified such data. This could increase the risk to patient data under private custodianship.

Conclusions

We are currently in a familiar situation in which regulation and oversight risk falling behind the technologies they govern. Regulation should emphasize patient agency and consent, and should encourage increasingly sophisticated methods of data anonymization and protection.

Rebooting consent in the digital age: a governance framework for health data exchange

In August 2020, India announced its vision for the National Digital Health Mission (NDHM), a federated national digital health exchange where digitised data generated by healthcare providers will be exported via application programme interfaces to the patient's electronic personal health record. The NDHM architecture is initially expected to be a claims platform for the national health insurance programme 'Ayushman Bharat' that serves 500 million people. Such large-scale digitisation and mobility of health data will have significant ramifications on care delivery, population health planning, as well as on the rights and privacy of individuals. Traditional mechanisms that seek to protect individual autonomy through patient consent will be inadequate in a digitised ecosystem where processed data can travel near instantaneously across various nodes in the system and be combined, aggregated, or even re-identified.In this paper we explore the limitations of 'informed' consent that is sought either when data are collected or when they are ported across the system. We examine the merits and limitations of proposed alternatives like the fiduciary framework that imposes accountability on those that use the data; privacy by design principles that rely on technological safeguards against abuse; or regulations. Our recommendations combine complementary approaches in light of the evolving jurisprudence in India and provide a generalisable framework for health data exchange that balances individual rights with advances in data science.

Ethical Issues in Patient Data Ownership

Patient data have conventionally been thought to be well protected by the privacy laws outlined in the United States. The increasing interest of for-profit companies in acquiring the databases of large health care systems poses new challenges to the protection of patients' privacy. It also raises ethical concerns of sharing patient data with entities that may exploit it for commercial interests and even target vulnerable populations. Recognizing that every breach in the confidentiality of large databases exposes millions of patients to the potential of being exploited is important in framing new rules for governing the sharing of patient data. Similarly, the ethical aspects of data voluntarily and altruistically provided by patients for research, which may be exploited for commercial interests due to patient data sharing between health care entities and third-party companies, need to be addressed. The rise of technologies such as artificial intelligence and the availability of personal data gleaned by data vendor companies place American patients at risk of being exploited both intentionally and inadvertently because of the sharing of their data by their health care provider institutions and third-party entities.

Overview and Limitations of Database Research in Anesthesiology: A Narrative Review

The utilization of large-scale databases for research in medical fields, including anesthesiology, has increased in popularity over the last decade, likely due to their structured content and relative ease of access. These databases have been used in a variety of perioperative studies, including analyses of risk stratification, preoperative testing, complications, and cost. While these databases contain a wealth of information that allows for an abundance of research opportunities, there are unique limitations to their use. A comprehensive understanding will afford the anesthesiology researcher the knowledge and tools to not only better interpret studies that utilized these databases, but also to conduct analyses of their own. This review details the content and composition of these databases, highlights the advantages of and limitations to their use, and offers information about their access and cost.

Internet of things issues related to psychiatry

Background

Internet of Things (IoT) devices for remote monitoring, diagnosis, and treatment are widely viewed as an important future direction for medicine, including for bipolar disorder and other mental illness. The number of smart, connected devices is expanding rapidly. IoT devices are being introduced in all aspects of everyday life, including devices in the home and wearables on the body. IoT devices are increasingly used in psychiatric research, and in the future may help to detect emotional reactions, mood states, stress, and cognitive abilities. This narrative review discusses some of the important fundamental issues related to the rapid growth of IoT devices.

Main body

Articles were searched between December 2019 and February 2020. Topics discussed include background on the growth of IoT, the security, safety and privacy issues related to IoT devices, and the new roles in the IoT economy for manufacturers, patients, and healthcare organizations.

Conclusions

The use of IoT devices will increase throughout psychiatry. The scale, complexity and passive nature of data collection with IoT devices presents unique challenges related to security, privacy and personal safety. While the IoT offers many potential benefits, there are risks associated with IoT devices, and from the connectivity between patients, healthcare providers, and device makers. Security, privacy and personal safety issues related to IoT devices are changing the roles of manufacturers, patients, physicians and healthcare IT organizations. Effective and safe use of IoT devices in psychiatry requires an understanding of these changes.

Ethical issues in using ambient intelligence in health-care settings

Ambient intelligence is increasingly finding applications in health-care settings, such as helping to ensure clinician and patient safety by monitoring staff compliance with clinical best practices or relieving staff of burdensome documentation tasks. Ambient intelligence involves using contactless sensors and contact-based wearable devices embedded in health-care settings to collect data (eg, imaging data of physical spaces, audio data, or body temperature), coupled with machine learning algorithms to efficiently and effectively interpret these data. Despite the promise of ambient intelligence to improve quality of care, the continuous collection of large amounts of sensor data in health-care settings presents ethical challenges, particularly in terms of privacy, data management, bias and fairness, and informed consent. Navigating these ethical issues is crucial not only for the success of individual uses, but for acceptance of the field as a whole.

Promises and perils of artificial intelligence in dentistry

Artificial intelligence (AI) is a subdiscipline of computer science that has made substantial progress in medicine and there is a growing body of AI research in dentistry. Dentists should have an understanding of the foundational concepts and the ability to critically evaluate dental research in AI. Machine learning (ML) is a subfield of AI that most dental AI research is dedicated to. The most prolific area of ML research is automated interpretation of dental imaging. Other areas include providing treatment recommendations, predicting future disease and treatment outcomes. The research impact is limited by small datasets that do not harness the positive correlation between very large datasets and ML performance. There is also a need to standardize research methodologies and utilize performance metrics that are appropriate for the clinical context. In addition to research challenges, this article discusses the ethical, legal and logistical considerations associated with implementation in clinical practice. This includes explainable AI, model bias, data privacy and security. The future implications of AI in dentistry involve a promise for a novel form of practicing dentistry however, the effect of AI on patient outcomes is yet to be determined.

Feasibility of Fitness Tracker Usage to Assess Activity Level and Toxicities in Patients With Colorectal Cancer

PURPOSE

Performance status (PS) is a subjective assessment of patients' overall health. Quantification of physical activity using a wearable tracker (Fitbit Charge [FC]) may provide an objective measure of patient's overall PS and treatment tolerance.

MATERIALS AND METHODS

Patients with colorectal cancer were prospectively enrolled into two cohorts (medical and surgical) and asked to wear FC for 4 days at baseline (start of new chemotherapy [± 4 weeks] or prior to curative resection) and follow-up (4 weeks [± 2 weeks] after initial assessment in medical and postoperative discharge in surgical cohort). Primary end point was feasibility, defined as 75% of patients wearing FC for at least 12 hours/d, 3 of 4 assigned days. Mean steps per day (SPD) were correlated with toxicities of interest (postoperative complication or ≥ grade 3 toxicity). A cutoff of 5,000 SPD was selected to compare outcomes.

RESULTS

Eighty patients were accrued over 3 years with 55% males and a median age of 59.5 years. Feasibility end point was met with 68 patients (85%) wearing FC more than predefined duration and majority (91%) finding its use acceptable. The mean SPD count for patients with PS 0 was 6,313, and for those with PS 1, it was 2,925 (122 and 54 active minutes, respectively) (P = .0003). Occurrence of toxicity of interest was lower among patients with SPD > 5,000 (7 of 33, 21%) compared with those with SPD < 5,000 (14 of 43, 32%), although not significant (P = .31).

CONCLUSION

Assessment of physical activity with FC is feasible in patients with colorectal cancer and well-accepted. SPD may serve as an adjunct to PS assessment and a possible tool to help predict toxicities, regardless of type of therapy. Future studies incorporating FC can standardize patient assessment and help identify vulnerable population.