Journal of Medical Systems最新文献

In this issue of Journal of Medical Systems, Neri et al. share results from their study in which they compared the YouCare device to a standard Holter monitor. The wearable used in the study incorporates a single electrocardiogram lead in a crop top garment that is customized for each patient. This editorial discusses the YouCare device, the study findings, and their clinical relevance and impact in the context of wearable technology.

This research evaluates the readability and quality of patient information material about female urinary incontinence (fUI) in ten popular artificial intelligence (AI) supported chatbots. We used the most recent versions of 10 widely-used chatbots, including OpenAI's GPT-4, Claude-3 Sonnet, Grok 1.5, Mistral Large 2, Google Palm 2, Meta's Llama 3, HuggingChat v0.8.4, Microsoft's Copilot, Gemini Advanced, and Perplexity. Prompts were created to generate texts about UI, stress type UI, urge type UI, and mix type UI. The modified Ensuring Quality Information for Patients (EQIP) technique and QUEST (Quality Evaluating Scoring Tool) were used to assess the quality, and the average of 8 well-known readability formulas, which is Average Reading Level Consensus (ARLC), were used to evaluate readability. When comparing the average scores, there were significant differences in the mean mQEIP and QUEST scores across ten chatbots (p = 0.049 and p = 0.018). Gemini received the greatest mean scores for mEQIP and QUEST, whereas Grok had the lowest values. The chatbots exhibited significant differences in mean ARLC, word count, and sentence count (p = 0.047, p = 0.001, and p = 0.001, respectively). For readability, Grok is the easiest to read, while Mistral is highly complex to understand. AI-supported chatbot technology needs to be improved in terms of readability and quality of patient information regarding female UI.

Stroke registries are tools for improving care and advancing research. We aim to describe the methodology and resourcing of existing national stroke registries. We conducted a systematic search of the published, peer-reviewed literature and grey literature examining descriptions of data collection methods and resourcing of national stroke registries published from 2012 to 2023. The systematic review was registered in PROSPERO (CRD42023393841). 101 records relating to 21 registries in 19 countries were identified. They universally employed web-based platforms for data collection. The principal profession of data collectors was nursing. All included the acute phase of care, 28% (6) registered the pre-hospital (ambulance) phase and 14% (3) included rehabilitation. 80% (17) collected outcome data. The registries varied in their approach to outcome data collection; in 9 registries it was collected by hospitals, in 2 it was collected by the registry, and 7 had linkage to national administrative databases allowing follow-up of a limited number of end points. Coverage of the total number of strokes varies from 6 to 95%. Despite widespread use of Electronic Health Records (EHRs) the ability to automatically populate variables remained limited. Governance and management structures are diverse, making it challenging to compare their resourcing. Data collection for clinical registries requires time and necessary skills and imposes a significant administrative burden on the professionals entering data. We highlight the role of clinical registries as powerful instruments for quality improvement. Future work should involve creating a central repository of stroke registries to enable the development of new registries and facilitate international collaboration.

In recent years a significant demand to develop computer-assisted diagnostic tools to assess prostate cancer using whole slide images has been observed. In this study we develop and validate a machine learning system for cancer assessment, inclusive of detection of perineural invasion and measurement of cancer portion to meet clinical reporting needs. The system analyses the whole slide image in three consecutive stages: tissue detection, classification, and slide level analysis. The whole slide image is divided into smaller regions (patches). The tissue detection stage relies upon traditional machine learning to identify WSI patches containing tissue, which are then further assessed at the classification stage where deep learning algorithms are employed to detect and classify cancer tissue. At the slide level analysis stage, entire slide level information is generated by aggregating all the patch level information of the slide. A total of 2340 haematoxylin and eosin stained slides were used to train and validate the system. A medical team consisting of 11 board certified pathologists with prostatic pathology subspeciality competences working independently in 4 different medical centres performed the annotations. Pixel-level annotation based on an agreed set of 10 annotation terms, determined based on medical relevance and prevalence, was created by the team. The system achieved an accuracy of 99.53% in tissue detection, with sensitivity and specificity respectively of 99.78% and 99.12%. The system achieved an accuracy of 92.80% in classifying tissue terms, with sensitivity and specificity respectively 92.61% and 99.25%, when 5x magnification level was used. For 10x magnification, these values were respectively 91.04%, 90.49%, and 99.07%. For 20x magnification they were 84.71%, 83.95%, 90.13%.

The worldwide nursing shortage has led to the exploration of using robotics to support care delivery and reduce nurses' workload. In this observational, mixed-method study, we examined the implementation of a robotic nurse assistant (RNA) in a hospital ward to support vital signs measurements, medication, and item delivery. Human-robot interaction was assessed in four domains: usability, social acceptance, user experience, and its societal impact. Patients in a general medicine ward were recruited to participate in a one-time trial with the RNA and a post-trial 75-question survey. Patients' interactions with the RNA were video recorded for analysis including patients' behaviours, facial emotions, and visual attention. Focus group discussions with nurses elicited their perceptions of working with the RNA, areas for improvement, and scalability. Sixty-seven patients aged 21-79 participated in the trial. Eight in 10 patients reported positive interactions with the RNA. When the RNA did not perform to expectations, only 25% of patients attributed fault to the RNA. Video analysis showed patients at ease interacting with the RNA despite some technical problems. Nurses saw potential for the RNA taking over routine tasks. However, they were sceptical of real time savings and were concerned with the RNA's ability to work well with older patients. Patients and nurses suggested greater interactivity between RNA and patients. Future studies should examine potential timesaving and whether time saved translated to nurses performing higher value clinical tasks. The utility of improved RNA's social capability in a hospital setting should be explored as well.

The exchange of medical images and patient data over the internet has attracted considerable attention in the past decade, driven by advancements in communication and health services. However, transferring confidential data through insecure channels, such as the internet, exposes it to potential manipulations and attacks. To ensure the authenticity of medical images while concealing patient data within them, this paper introduces a high-capacity and reversible fragile watermarking model in which an authentication watermark is initially generated from the cover image and merged with the patient's information, photo, and medical report to form the global watermark. This watermark is subsequently encrypted using the chaotic Chen system technique, enhancing the model's security and ensuring patient data confidentiality. The cover image then undergoes a Discrete Fourier Transform (DFT) and the encrypted watermark is inserted into the frequency coefficients using a new embedding technique. The experimental results demonstrate that the proposed method achieves great watermarked image quality, with a PSNR exceeding 113 dB and an SSIM close to 1, while maintaining a high embedding capacity of 3 BPP (Bits Per Pixel) and offering perfect reversibility. Furthermore, the proposed model demonstrates high sensitivity to attacks, successfully detecting tampering in all 18 tested attacks, and achieves nearly perfect watermark extraction accuracy, with a Bit Error Rate (BER) of 0.0004%. This high watermark extraction accuracy is crucial in our situation where patient data need to be retrieved from the watermarked images with almost no alteration.

The use of artificial intelligence (AI) in the segmentation of liver structures in medical images has become a popular research focus in the past half-decade. The performance of AI tools in screening for this task may vary widely and has been tested in the literature in various datasets. However, no scientometric report has provided a systematic overview of this scientific area. This article presents a systematic and bibliometric review of recent advances in neuronal network modeling approaches, mainly of deep learning, to outline the multiple research directions of the field in terms of algorithmic features. Therefore, a detailed systematic review of the most relevant publications addressing fully automatic semantic segmenting liver structures in Computed Tomography (CT) images in terms of algorithm modeling objective, performance benchmark, and model complexity is provided. The review suggests that fully automatic hybrid 2D and 3D networks are the top performers in the semantic segmentation of the liver. In the case of liver tumor and vasculature segmentation, fully automatic generative approaches perform best. However, the reported performance benchmark indicates that there is still much to be improved in segmenting such small structures in high-resolution abdominal CT scans.

Data on the health of transgender and gender diverse (TGD) people are scarce. Researchers are increasingly turning to insurance claims data to investigate disease burden among TGD people. Since claims do not include gender self-identification or modality (i.e., TGD or not), researchers have developed algorithms to attempt to identify TGD individuals using diagnosis, procedure, and prescription codes, sometimes also inferring sex assigned at birth and gender. Claims-based algorithms introduce epistemological and ethical complexities that have yet to be addressed in data informatics, epidemiology, or health services research. We discuss the implications of claims-based algorithms to identify and categorize TGD populations, including perpetuating cisnormative biases and dismissing TGD individuals' self-identification. Using the framework of epistemic injustice, we outline ethical considerations when undertaking claims-based TGD health research and provide suggestions to minimize harms and maximize benefits to TGD individuals and communities.

Administering medications to patients with documented drug hypersensitivity reactions (DHR) poses a significant risk for adverse events, ranging from mild reactions to life-threatening incidents. Electronic healthcare systems have revolutionized the modern clinical decision-making process, with built in warnings. However, as these alerts become a routine part of healthcare provider's workflow, alert fatigue becomes a challenge. This study was conducted within the Ministry of National Guard Health Affairs (MNGHA), a government healthcare system in Saudi Arabia. A taskforce of experts was formed to develop an electronic path that would prevent unintentional overrides of severe drug allergy alerts. The system underwent rigorous testing, and monitoring parameters were established. We outline the implementation of a system upgrade designed to trigger an alternative interruption in the computerized physician order entry (CPOE) process, distinct from the regular allergy pop-up alerts. The alternate path is activated upon a CPOE with a drug-to-drug match and a documented severe drug allergy symptom, necessitating co-signature form another prescriber before proceeding. The adopted upgrade is a proactive approach to enhance medication safety in electronic healthcare systems, ensuring that serious allergy-related warnings are not overridden, ultimately enhancing patient safety. Further monitoring will confirm the safety and effectiveness of this measure. This study provides a model for institutions seeking to prevent allergy-related harm within their patient population.