Patient Privacy Research Articles

Facilitators and barriers to AI adoption in nursing practice: a qualitative study of registered nurses' perspectives

BackgroundIntegrating Artificial Intelligence (AI) in nursing practice is revolutionising healthcare by enhancing clinical decision-making and patient care. However, the adoption of AI by registered nurses, especially in varied healthcare settings such as Saudi Arabia, remains underexplored. Understanding the facilitators and barriers from the perspective of frontline nurses is crucial for successful AI implementation.AimThis study aimed to explore registered nurses' perspectives on the facilitators and barriers to AI adoption in nursing practice in Saudi Arabia and to propose an extended Technology Acceptance Model for AI in Nursing (TAM-AIN).MethodsA qualitative study utilising focus group discussions was conducted with 48 registered nurses from four major healthcare facilities in Al-Kharj, Saudi Arabia. Thematic analysis, guided by the Technology Acceptance Model framework, was employed to analyse the data.ResultsKey facilitators of AI adoption included perceived benefits to patient care (85%), strong organisational support (70%), and comprehensive training programs (75%). Primary barriers involved technical challenges (60%), ethical concerns regarding patient privacy (55%), and fears of job displacement (45%). These findings led to the development of TAM-AIN, an extended model that incorporates additional constructs such as ethical alignment, organisational readiness, and perceived threats to professional autonomy.ConclusionsAI adoption in nursing practice requires a holistic approach that addresses technical, educational, ethical, and organisational challenges. The proposed TAM-AIN offers a comprehensive framework for optimising AI integration into nursing practice, emphasising the importance of nurse-centred implementation strategies. This model provides healthcare institutions and policymakers with a robust tool to facilitate successful AI adoption and enhance patient outcomes.

GAN-Based Novel Approach for Generating Synthetic Medical Tabular Data

The generation of synthetic medical data has become a focal point for researchers, driven by the increasing demand for privacy-preserving solutions. While existing generative methods heavily rely on real datasets for training, access to such data is often restricted. In contrast, statistical information about these datasets is more readily available, yet current methods struggle to generate tabular data solely from statistical inputs. This study addresses the gaps by introducing a novel approach that converts statistical data into tabular datasets using a modified Generative Adversarial Network (GAN) architecture. A custom loss function was incorporated into the training process to enhance the quality of the generated data. The proposed method is evaluated using fidelity and utility metrics, achieving “Good” similarity and “Excellent” utility scores. While the generated data may not fully replace real databases, it demonstrates satisfactory performance for training machine-learning algorithms. This work provides a promising solution for synthetic data generation when real datasets are inaccessible, with potential applications in medical data privacy and beyond.

Traversable Ledger for Responsible Data Sharing and Access Control in Health Research

Healthcare institutions and health registries often store patients’ health data. In order to ensure privacy, sensitive medical information is stored separately from the identifying information of the patient. Generally, institutions anonymize medical information while sharing it for external use. However, internal users may also use it for identifying inaccuracies or missing information. Even though internal users may be legally permitted to access sensitive medical information, such access may lead to the identification of the patient, which can be vulnerable to patient privacy. Ensuring the accountability and responsibility of the internal users may lead to tractability in case of adversarial access with malicious intentions. Therefore, a secure system must be developed for the storage and retrieval of health data. To this end, in this paper, we propose a ledger-based system that cryptographically ensures that all access to health data must be logged into a ledger. Nevertheless, the ledger entries must be protected against adversarial access, too. At the same time, the ledger must be traversable by the patients as well as internal users. To address these needs, we propose techniques for the construction of a ledger to permit both internal users and patients to securely traverse and view only the entries to which they are linked.

Trends, prospects, challenges, and security in the healthcare internet of things

The Healthcare Internet of Things (H-IoT) is a rapidly developing problem solving model with significant potential to improve patient care and healthcare outcomes. This study focuses on integrating cryptographic platforms into H-IoT systems to enable secure data access. We present insights on how to manage challenges such as cyber risks, resource constraints, latency, and energy consumption by exploring cryptographic approaches in diverse H-IoT applications. We explore important topics including big data management, blockchain, machine learning, edge computing, and software-defined networks. Additionally, we examine real-time operational challenges and emerging trends, such as remote patient monitoring and predictive analytics. Our research underscores the need for strong encryption mechanisms, access controls, device authentication, and proactive threat detection to improve the safety and efficiency of healthcare services. By combining cryptographic approaches with the architecture of the H-IoT system, this work provides the foundation for resilient healthcare infrastructures. Addressing current challenges and anticipating future opportunities contribute to strategic healthcare planning that prioritizes patient privacy and data security while anticipating future strategic healthcare planning opportunities. This article aims to provide valuable information to researchers by covering energy-efficient and resource-optimized healthcare system approaches that are integrated into the development of H-IoT systems.

Python technology and its applications in radiomics

Python, developed by Guido van Rossum, is favored for its simplicity and extensive ecosystem of libraries, which facilitate efficient coding and integration with other programming languages. Here, we aim to explore and summarize the role of Python in radiomics, a field focused on extracting and analyzing quantitative features from medical imaging to improve disease characterization and treatment evaluation. Radiomics addresses the complexities of tumor heterogeneity by transforming imaging data from modalities such as computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET) into actionable insights, often using statistical methods and machine learning techniques. Its primary applications include differentiating between benign and malignant tumors and predicting treatment outcomes, etc. Python is integral to several stages of radiomics, including image acquisition, region of interest (ROI) segmentation, feature extraction, and statistical analysis. By utilizing libraries such as PyRadiomics and Scikit-learn, researchers can significantly enhance the accuracy and efficiency of their analyses. Looking forward, Python holds considerable promise in radiomics, especially with ongoing advancements in medical imaging and big data. However, challenges such as data standardization, model interpretability, and patient privacy protection must be addressed to fully unlock its potential for improving diagnostic precision and patient outcomes.

Introduction to Generative Artificial Intelligence: Contextualizing the Future.

Generative artificial intelligence (GAI) is a promising new technology with the potential to transform communication and workflows in health care and pathology. Although new technologies offer advantages, they also come with risks that users, particularly early adopters, must recognize. Given the fast pace of GAI developments, pathologists may find it challenging to stay current with the terminology, technical underpinnings, and latest advancements. Building this knowledge base will enable pathologists to grasp the potential risks and impacts that GAI may have on the future practice of pathology. To present key elements of GAI development, evaluation, and implementation in a way that is accessible to pathologists and relevant to laboratory applications. Information was gathered from recent studies and reviews from PubMed and arXiv. GAI offers many potential benefits for practicing pathologists. However, the use of GAI in clinical practice requires rigorous oversight and continuous refinement to fully realize its potential and mitigate inherent risks. The performance of GAI is highly dependent on the quality and diversity of the training and fine-tuning data, which can also propagate biases if not carefully managed. Ethical concerns, particularly regarding patient privacy and autonomy, must be addressed to ensure responsible use. By harnessing these emergent technologies, pathologists will be well placed to continue forward as leaders in diagnostic medicine.

Emerging Educational and Networking Platforms in Interventional Cardiology: Expanding Beyond Traditional Social Media.

With ubiquitous use of smartphones and other Internet-enabled devices, clinicians have access to a plethora of multimedia platforms that enable them to network, share educational content, and learn. Several platforms also enable resources that support clinical decision making. Traditional social media has been used extensively by cardiologists for several years and now extends to the entire care team including trainees, advanced practitioners, nurses and technologists. Numerous challenges include concerns about patient privacy and security, lack of peer-review and organization, and chances of propagating misinformation. More recently, newer networking platforms have emerged within cardiology to overcome several of the limitations of traditional social media. Additionally, mobile applications have emerged as an educational tool that provides clinicians with easy access to the latest information including guidelines, consensus documents, risk estimators and treatment algorithms. These point-of-care learning and decision-making tools are especially relevant and impactful in a procedural field like interventional cardiology. While professional societies are major drivers of these emerging platforms, there are also other stakeholders including educational institutions and the medical device industry. In this review article, we examine the role of traditional social media and describe newer networking platforms as well as mobile applications applicable to interventional cardiology with a discussion about their relative advantages and disadvantages.

Cybersecurity Threats in Healthcare IT: Challenges, Risks, and Mitigation Strategies

The healthcare industry's reliance on digital technology to improve patient care, operations, and performance has exposed it to cybersecurity threats. The extensive capture, storage, and transmission of confidential data by healthcare institutions has led to a significant prevalence of cyber-attacks. The intersection of healthcare and information technology presents considerable challenges, with patient privacy, the security of medical devices, and the continuity of essential healthcare services all at risk. This study offers a contemporary contextual analysis of IT cybersecurity concerns within the healthcare sector. It employs a systematic methodology to delineate critical risk factors and threats, the strategies they might utilize to attain their compromise aims, and the repercussions they may impose on healthcare delivery. It also explores the measures and guidelines to address this risk. The study highlights the existence of many threats experienced by healthcare organizations, such as data theft and leakage, ransomware, cyber-attacks in different forms, and contaminated medical equipment. One prominent example is the WannaCry attack on the United Kingdom's National Health Service in 2017. According to the survey, there is a severe scarcity of qualified personnel in healthcare organizations who can defend against cyber threats effectively and efficiently, underscoring the urgent need for skilled professionals in cybersecurity. This analysis also discusses various considerations concerning achieving high availability and protecting data in healthcare businesses. It examines how emerging applications like IoMT enhance the quality of medical services for patients and the security challenges these devices pose. The report evaluates the sufficiency of existing regulatory guidelines in addressing emerging dangers, including HIPAA vulnerabilities. The current definitive framework underscores the critical importance of enhancing cybersecurity within healthcare IT. It elucidates optimal strategies for mitigating cyber hazards by promoting technological prevention, staff training, and adherence to organizational policies. The work emphasizes the necessity for ongoing research, the establishment of collaboration between medical professionals and IT security experts, and the formulation of effective strategies to protect patient information and ensure the continuity of healthcare delivery in the face of increasing cyber-attack threats in digitalization.

Perspectives of Patients Regarding Artificial Intelligence and Its Application in Healthcare: A Qualitative Study.

Artificial intelligence integration into healthcare has gained significant attention in recent years, with its use ranging from disease diagnosis to surgical assistance. While artificial intelligence's potential to improve patient outcomes and optimise patient care is undeniable, concerns regarding privacy, transparency, and the potential for medical errors persist. To take full advantage of artificial intelligence's transformative abilities, understanding patient perceptions and attitudes towards its integration into medicine is crucial for ethical considerations and health outcomes. This study aimed to describe patients' perceptions of medical artificial intelligence and its integration into the healthcare system, drawing attention to a crucial yet understudied aspect of artificial intelligence adoption in Kazakhstan. Descriptive qualitative design. From February to March 2024, the researchers conducted semi-structured interviews amongst 13 patients. The interviews were audio-recorded, transcribed, translated, and then analysed using a thematic analysis approach. The study adhered to the COREQ guidelines. Five themes emerged from 13 interviews: the benefits of artificial intelligence on patient care, the importance of human factors on patient care over artificial intelligence, revolutionising patient care delivery through artificial intelligence, patient education and artificial intelligence, and balancing technology and human interaction in artificial intelligence-driven intervention. Patient perceptions of artificial intelligence integration into healthcare were primarily positive. Nevertheless, patients prefer artificial intelligence as a supplementary tool under human supervision due to risks such as possible medical errors and violations of patient privacy. Patients provided the data for this study. The researchers interviewed them about their perceptions of medical artificial intelligence and its integration into the healthcare system. The patients or the public contributed nothing to the other aspects of the study.

Deep Learning-Based Real Time Human Detection System Using LiDAR Data for Smart Healthcare Monitoring

Abstract Continuous patient monitoring is a critical component in healthcare systems to ensure patient safety and well-being. Traditionally, this monitoring requires significant oversight by healthcare professionals, making it resourceintensive. Common methods, such as cameras, often compromise patient privacy by capturing detailed visual information. LiDAR technology presents a non-intrusive alternative for continuous monitoring, offering effectiveness across various lighting conditions due to its use of laser light. It provides high-resolution depth information encoded as intensity, making it an ideal solution for maintaining patient privacy while delivering precise monitoring capabilities. LiDAR was selected as a reference method because of its precision and ability to capture detailed depth information, which are crucial for accurate monitoring. In this project, a solution that leverages LiDAR technology to enhance patient monitoring systems was introduced. By training a YOLOv5 deep learning model using transfer learning, a method for accurate human detection and tracking within rooms using data collected from a digital LiDAR sensor was used. The high accuracy of the LiDAR sensor enables precise human tracking, which is critical for timely interventions and ensuring patient safety. This approach accelerates response times, preserves privacy, ensures safety during disease outbreaks, addresses healthcare worker shortages, and facilitates efficient monitoring of multiple patients simultaneously without invasive sensors. The ultimate goal of this project is to integrate this solution into Ambient Assisted Living (AAL) systems for elderly individuals, providing them with a safer and more autonomous living environment. Our trained YOLOv5 model has demonstrated exceptional performance, achieving a Mean Average Precision (mAP) of 99.4% at an Intersection over Union (IoU) threshold of 0.5, with a Precision of 99.6% and a Recall of 99.4%.

Recent Advances in Pharmaceutical Design: Unleashing the Potential of Novel Therapeutics.

Pharmaceutical design has made significant advancements in recent years, leading to the development of novel therapeutics with unprecedented efficacy and safety profiles. This review highlights the potential of these innovations to revolutionize healthcare and improve patient outcomes. The application of cutting-edge technologies like artificial intelligence, machine learning, and data mining in drug discovery and design has made it easier to find potential drug candidates. Combining big data and omics has led to the discovery of new therapeutic targets and personalized medicine strategies. Nanoparticles, liposomes, and microneedles are examples of advanced drug delivery systems that allow precise control over drug release, better bioavailability, and targeted delivery to specific tissues or cells. This improves the effectiveness of the treatment while reducing side effects. Stimuli-responsive materials and smart drug delivery systems enable drugs to be released on demand when specific internal or external signals are sent. Biologics and gene therapies are promising approaches in pharmaceutical design, offering high specificity and potency for treating various diseases like cancer, autoimmune disorders, and infectious diseases. Gene therapies hold tremendous potential for correcting genetic abnormalities, with recent breakthroughs demonstrating successful outcomes in inherited disorders and certain types of cancer. Advancements in nanotechnology and nanomedicine have paved the way for innovative diagnostic tools and therapeutics, such as nanoparticle-based imaging agents, targeted drug delivery systems, gene editing technologies, and regenerative medicine strategies. Finally, the review emphasizes the importance of regulatory considerations, ethical challenges, and future directions in pharmaceutical design. Regulatory agencies are adapting to the rapid advancements in the field, ensuring the safety and efficacy of novel therapeutics while fostering innovation. Ethical considerations regarding the use of emerging technologies, patient privacy, and access to advanced therapies also require careful attention.

Harmonizing EU Health Legislation for New Genomic Techniques

This article examines the regulation and application of New Genomic Techniques (NGTs) and genetic testing within the European Union (EU). It scrutinises the detailed aspects of EU data protection laws, underscoring their significance in safeguarding patient privacy and facilitating data sharing in healthcare and research sectors. The paper delves into the ethical dimensions and technological advances in genetic engineering, specifically focusing on gene editing and direct-to-consumer genetic testing, illuminating the complex interplay between technology, ethics, and legal frameworks. The analysis extends to the diverse regulatory environments at both national and EU levels, highlighting the urgent need for legislative updates in response to the rapid advancements in genomic technologies. This necessity accentuates the ongoing challenge of balancing fostering innovation and adhering to ethical and legal standards. Moreover, the article addresses the efforts to harmonise genetic testing regulations across European nations, considering the perspectives of multiple stakeholders. The discourse presented in the article underscores the complexity and critical importance of formulating policies that safeguard public interests and propel scientific advancements in this swiftly evolving domain. The EU should enhance the regulation and standardisation of NGTs in the field of human health and medicine by updating legislation, implementing risk assessment and safety protocols, establishing ethical guidelines, involving various stakeholders, adhering to global standards, promoting public education, and integrating NGTs into healthcare systems, ultimately aiming to promote safe and ethical NGTs development and advance public health objectives.

Re-identification of anonymised MRI head images with publicly available software: investigation of the current risk to patient privacy

Facial recognition software (FRS) has historically been perceived as lacking the capability to identify individuals from cross-sectional medical images. Utilising such data for identification purposes was considered infeasible due to the substantial computational power and specialised technical expertise it would require. However, recent advancements in accessible artificial intelligence-based (AI-based) software and open-source tools have made these applications widely available and easy to use, raising new privacy concerns. This proof-of-concept was designed as a cross-sectional study and included participants with a verified online presence. Standard magnetic resonance imaging (MRI) head scans were performed on these participants, from which three-dimensional rendering (3DR) images were created using free and publicly available software. These images were used for face searches by free and publicly available FRS. Different head orientations and hairstyles were applied to the 3DR images to assess whether non-facial features influenced the FRS results. All results were obtained between the 10th of February 2024 and the 1st of March2024. Face searches of 3DR images in a database containing over 800 million images from the World Wide Web (WWW) yielded correct matches for 50% of the participants in less than 10min. The user-friendly software required minimal computational knowledge or resources, making this process broadly accessible. Modifying elements such as hairstyles or the orientation of the 3DR to better resemble actual photographs of the participants improved FRS matches. Current existing FRS can swiftly and accurately identify individuals from MRI head scans. This poses a significant privacy risk for participants in enrolled clinical trials and highlights the urgent need for improved data protection measures and increased sensitivity to ensure participant confidentiality. There was no funding source for this study.

Artificial intelligence in ophthalmology.

To review role of artificial intelligence in medicine. Artificial intelligence is continuing to revolutionize access, diagnosis, personalization of medicine, and treatment in healthcare. As a matter of fact, artificial intelligence contributed to the research that resulted in 2024 Nobel Prizes in physics, chemistry, and economics. We are only at the tip of the iceberg in utilizing the abilities of artificial intelligence in medicine to improve accuracy of diagnoses and to enhance patient outcomes. Artificial intelligence has allowed better image analysis, prediction of progression of disease, personalized treatment plans, incorporations of genomics, and improved efficiency in care and follow-up utilizing home monitoring. In ocular health diagnosis and treatment of diabetic retinopathy, macular degeneration, glaucoma, corneal infections, and ectasia are only a few examples of how the power of artificial intelligence has been harnessed. Even though there are still challenges that need more work in the areas of patient privacy, Health Insurance Portability and Accountability Act (HIPAA) compliance, reliability, and development of regulatory frameworks, artificial intelligence has revolutionized and will continue to revolutionize medicine. Artificial intelligence is enhancing medical diagnosis and treatment, as well as access and prevention. Ocular imaging, visual outcome, optics, intraocular pressure, and data points will continue to see growth it the field of artificial intelligence.

AI at the Edge: Enhancing Cybersecurity and Real-Time Decision-Making in Healthcare IoT Systems

The integration of Artificial Intelligence (AI) with edge computing and Internet of Things (IoT) systems has revolutionized healthcare, enabling real-time decision-making and personalized care delivery. However, these advancements also introduce significant cybersecurity challenges that threaten patient privacy and system integrity. This research explores the convergence of AI, edge computing, and IoT in healthcare, focusing on enhancing real-time decision-making while mitigating cybersecurity risks. By leveraging edge computing’s decentralized architecture, this study aims to reduce latency, optimize resource utilization, and ensure data security during transmission and processing. Machine learning models at the edge facilitate real-time analytics and decision-making, while robust cybersecurity frameworks, including encryption, anomaly detection, and secure communication protocols, are proposed to safeguard sensitive healthcare data. The research also examines the ethical implications and scalability of deploying secure AI-driven IoT systems in healthcare. By addressing these challenges, the study aims to create a resilient infrastructure for delivering efficient, secure, and reliable healthcare solutions.

Advancing Healthcare: Intelligent Speech Technology for Transcription, Disease Diagnosis, and Interactive Control of Medical Equipment in Smart Hospitals

Intelligent Speech Technology (IST) is revolutionizing healthcare by enhancing transcription accuracy, disease diagnosis, and medical equipment control in smart hospital environments. This study introduces an innovative approach employing federated learning with Multi-Layer Perceptron (MLP) and Gated Recurrent Unit (GRU) neural networks to improve IST performance. Leveraging the “Medical Speech, Transcription, and Intent” dataset from Kaggle, comprising a variety of speech recordings and corresponding medical symptom labels, noise reduction was applied using a Wiener filter to improve audio quality. Feature extraction through MLP and sequence classification with GRU highlighted the model’s robustness and capacity for detailed medical understanding. The federated learning framework enabled collaborative model training across multiple hospital sites, preserving patient privacy by avoiding raw data exchange. This distributed approach allowed the model to learn from diverse, real-world data while ensuring compliance with strict data protection standards. Through rigorous five-fold cross-validation, the proposed Fed MLP-GRU model demonstrated an accuracy of 98.6%, with consistently high sensitivity and specificity, highlighting its reliable generalization across multiple test conditions. In real-time applications, the model effectively performed medical transcription, provided symptom-based diagnostic insights, and facilitated hands-free control of healthcare equipment, reducing contamination risks and enhancing workflow efficiency. These findings indicate that IST, powered by federated neural networks, can significantly improve healthcare delivery, accuracy in patient diagnosis, and operational efficiency in clinical settings. This research underscores the transformative potential of federated learning and advanced neural networks for addressing pressing challenges in modern healthcare and setting the stage for future innovations in intelligent medical technology.

Generative AI in Improving Personalized Patient Care Plans: Opportunities and Barriers Towards Its Wider Adoption

The main aim of this study is to investigate the opportunities, challenges, and barriers in implementing generative artificial intelligence (Gen AI) in personalized patient care plans (PPCPs). This systematic review paper provides a comprehensive analysis of the current state, potential applications, and opportunities of Gen AI in patient care settings. This review aims to serve as a key resource for various stakeholders such as researchers, medical professionals, and data governance. We adopted the PRISMA review methodology and screened a total of 247 articles. After considering the eligibility and selection criteria, we selected 13 articles published between 2021 and 2024 (inclusive). The selection criteria were based on the inclusion of studies that report on the opportunities and challenges in improving PPCPs using Gen AI. We found that a holistic approach is required involving strategy, communications, integrations, and collaboration between AI developers, healthcare professionals, regulatory bodies, and patients. Developing frameworks that prioritize ethical considerations, patient privacy, and model transparency is crucial for the responsible deployment of Gen AI in healthcare. Balancing these opportunities and challenges requires collaboration between wider stakeholders to create a robust framework that maximizes the benefits of Gen AI in healthcare while addressing the key challenges and barriers such as explainability of the models, validation, regulation, and privacy integration with the existing clinical workflows.

A qualitative analysis of the use of intraoperative acupuncture for patients with nosocomephobia.

Nosocomephobia, a type of posttraumatic stress disorder (PTSD), is a diagnosis of an extreme fear of hospitals that can hinder current/future medical care. There is little research on how nosocomephobia affects elective surgery or how acupuncture can help patients cope. Using the transactional model of stress/coping, this qualitative case study examined the role of acupuncture in nosocomephobia patients' elective surgery appraisal processes. Two patients were interviewed about their nosocomephobia and prior hospital experiences. Six reviewers coded interview transcripts line-by-line using Dedoose software. Reviewers labeled meaningful words, phrases and sentences and produced over 600 codes. Reviewers discussed/identified themes by grouping similar codes and resolving discrepancies. A thematic analysis was then used to develop final themes. Pseudonyms were assigned to protect patient privacy. Sophie had avascular necrosis in both hips and suffered PTSD from a previous traumatic event. Intraoperative acupuncture calmed her hospital anxiety, allowing her to have both hips replaced. Olivia had experienced PTSD and hospital phobia since she was 12 years old. Acupuncture reduced her anxiety surrounding a necessary knee arthrotomy and osteochondral allograft. Thematic analysis showed how nosocomephobia impacted patients' views of surgery and distinguished between their unique fear rationale. The transactional model of stress/coping illustrated patients' appraisal process from surgery (stressor) to coping (acupuncture) to reappraisal (mental state). Procedural visits can be stressful due to already heightened anxiety. Although no definitive conclusions can be drawn from this small, uncontrolled case series, acupuncture may represent a safe, noninvasive way for nosocomephobia patients to manage preoperative anxiety.

Large-scale long-tailed disease diagnosis on radiology images

Developing a generalist radiology diagnosis system can greatly enhance clinical diagnostics. In this paper, we introduce RadDiag, a foundational model supporting 2D and 3D inputs across various modalities and anatomies, using a transformer-based fusion module for comprehensive disease diagnosis. Due to patient privacy concerns and the lack of large-scale radiology diagnosis datasets, we utilize high-quality, clinician-reviewed radiological images available online with diagnosis labels. Our dataset, RP3D-DiagDS, contains 40,936 cases with 195,010 scans covering 5568 disorders (930 unique ICD-10-CM codes). Experimentally, our RadDiag achieves 95.14% AUC on internal evaluation with the knowledge-enhancement strategy. Additionally, RadDiag can be zero-shot applied or fine-tuned to external diagnosis datasets sourced from various medical centers, demonstrating state-of-the-art results. In conclusion, we show that publicly shared medical data on the Internet is a tremendous and valuable resource that can potentially support building strong models for image understanding in healthcare.

Anonymize or synthesize? Privacy-preserving methods for heart failure score analytics

Abstract Aims Data availability remains a critical challenge in modern, data-driven medical research. Due to the sensitive nature of patient health records, they are rightfully subject to stringent privacy protection measures. One way to overcome these restrictions is to preserve patient privacy by using anonymization and synthetization strategies. In this work, we investigate the effectiveness of these methods for protecting patient privacy using real-world cardiology health records. Methods and results We implemented anonymization and synthetization techniques for a structure data set, which was collected during the HiGHmed Use Case Cardiology study. We employed the data anonymization tool ARX and the data synthetization framework ASyH individually and in combination. We evaluated the utility and shortcomings of the different approaches by statistical analyses and privacy risk assessments. Data utility was assessed by computing two heart failure risk scores on the protected data sets. We observed only minimal deviations to scores from the original data set. Additionally, we performed a re-identification risk analysis and found only minor residual risks for common types of privacy threats. Conclusion We could demonstrate that anonymization and synthetization methods protect privacy while retaining data utility for heart failure risk assessment. Both approaches and a combination thereof introduce only minimal deviations from the original data set over all features. While data synthesis techniques produce any number of new records, data anonymization techniques offer more formal privacy guarantees. Consequently, data synthesis on anonymized data further enhances privacy protection with little impacting data utility. We share all generated data sets with the scientific community through a use and access agreement.