Background: The rapid emergence of artificial intelligence–based large language models (LLMs) in 2022 has initiated extensive discussions within the academic community. While proponents highlight LLMs’ potential to improve writing and analytical tasks, critics caution against the ethical and cultural implications of widespread reliance on these models. Existing literature has explored various aspects of LLMs, including their integration, performance, and utility, yet there is a gap in understanding the nature of these discussions and how public perception contrasts with expert opinion in the field of public health. Objective: This study sought to explore how the general public’s views and sentiments regarding LLMs, using OpenAI’s ChatGPT as an example, differ from those of academic researchers and experts in the field, with the goal of gaining a more comprehensive understanding of the future role of LLMs in health care. Methods: We used a hybrid sentiment analysis approach, integrating the Syuzhet package in R (R Core Team) with GPT-3.5, achieving an 84% accuracy rate in sentiment classification. Also, structural topic modeling was applied to identify and analyze 8 key discussion topics, capturing both optimistic and critical perspectives on LLMs. Results: Findings revealed a predominantly positive sentiment toward LLM integration in health care, particularly in areas such as patient care and clinical decision-making. However, concerns were raised regarding their suitability for mental health support and patient communication, highlighting potential limitations and ethical challenges. Conclusions: This study underscores the transformative potential of LLMs in public health while emphasizing the need to address ethical and practical concerns. By comparing public discourse with academic perspectives, our findings contribute to the ongoing scholarly debate on the opportunities and risks associated with LLM adoption in health care.

Introduction: Atrial fibrillation (AF) in end-stage kidney disease (ESKD) and kidney transplant (KTx) recipients presents challenges in stroke risk management. This study aimed to compare hospitalization rates for ischemic and hemorrhagic cerebrovascular events in ESKD and KTx patients with and without AF. Methods: Using the National Inpatient Sample (2005–2019), retrospective analysis was conducted on hospitalizations for ESKD and KTx patients with and without AF. Baseline characteristics and hospitalization rates for five cerebral ischemic conditions and one hemorrhagic condition were compared. Descriptive statistics and t-tests were employed for analysis. Results: Among ESKD patients, those with AF exhibited significantly higher hospitalization rates for ischemic stroke, including 1)Cerebral infarction due to thrombosis, embolism, occlusion (0.11% vs. 0.08%,p<0.001), 2)Cerebral infarction due to thrombosis, embolism, and unspecified occlusion (1.93% vs. 1.51%, p<0.001), 3)Artery occlusion resulting in cerebral ischemia (1.37% vs. 0.93%,p<0.001), 4)Cerebral artery occlusion resulting in cerebral ischemia (0.48% vs. 0.42%,p<0.001), while experiencing lower rates of intraoperative and postprocedural cerebrovascular infarction (0.88% vs. 0.97%,p<0.001) compared to those without AF. Conversely, KTx patients with AF showed increased hospitalizations for hemorrhagic stroke, particularly nontraumatic intracranial hemorrhage (0.79% vs. 0.56%,p<0.001), compared to those without AF. However, they did not exhibit significant differences in hospitalization rates for most ischemic conditions, except for cerebral infarction due to thrombosis, embolism, and unspecific occlusion (1.62% vs. 1.11%,p<0.001) and artery occlusion resulting in cerebral ischemia (0.84% vs. 0.52%,p<0.001). Conclusion: Our findings reveal patterns in hospitalization rates between ESKD and KTx patients with AF compared to those without AF, with ESKD patients with AF exhibiting higher rates of ischemic stroke compared to ESKD patients without AF and KTx patients with AF showing increased hospitalizations for hemorrhagic stroke compared to those without AF. These findings demonstrate the impact of AF on hospitalization rates for ischemic and hemorrhagic cerebrovascular events in both ESKD and KTx patients.

In ultra-low light conditions the presence of dark current becomes a major source of noise for a CMOS sensor. Standard dark current compensation techniques, such as using a dark reference frame, bring significant improvements to dark noise in typical applications. However, applications requiring long integration times mean that such techniques cannot always be used. This paper presents a differential dark current compensating pixel. The pixel is made up of a differential amplifier and two photodiodes: one light shielded photodiode connected to the non-inverting input of the opamp and a light detecting photodiode connected to the inverting input of the opamp. An integrating capacitor is used in the feedback loop to convert photocurrent to voltage, and a switched capacitor network is present in parallel with the light shielded pixel, which is used to satisfy the output equation to compensate the dark current. The pixel uses 150 μm x 150 μm photodiodes and is fabricated in a standard 0.18 μm, 6M1P, CMOS process. The results show that the pixel is light sensitive and has a linear output as expected. However, the dark current is not predictably controlled. Further work will be carried out on the pixel design, and particularly the switched capacitor circuit, to determine the cause of the non-predictability of the pixel output.

Individuals with end-stage kidney disease (ESKD) face higher cerebrovascular risk. Yet, the impact of peripheral vascular disease (PVD) and kidney transplantation (KTx) on hospitalization rates for cerebral infarction and hemorrhage remains underexplored. Analyzing 2,713,194 ESKD hospitalizations (2005–2019) using the National Inpatient Sample, we investigated hospitalization rates for ischemic and hemorrhagic cerebrovascular diseases concerning ESKD, PVD, KTx, or their combinations. Patients hospitalized with cerebral infarction due to thrombosis/embolism/occlusion (CITO) or artery occlusion resulting in cerebral ischemia (AOSI) had higher rates of comorbid ESKD and PVD (4.17% and 7.29%, respectively) versus non-CITO or AOSI hospitalizations (2.34%, p < 0.001; 2.29%, p < 0.001). Conversely, patients hospitalized with nontraumatic intracranial hemorrhage (NIH) had significantly lower rates of ESKD and PVD (1.64%) compared to non-NIH hospitalizations (2.34%, p < 0.001). Furthermore, hospitalizations for CITO or AOSI exhibited higher rates of KTx and PVD (0.17%, 0.09%, respectively) compared to non-CITO or AOSI hospitalizations (0.05%, p = 0.033; 0.05%, p = 0.002). Patients hospitalized with NIH showed similar rates of KTx and PVD (0.04%) versus non-NIH hospitalizations (0.05%, p = 0.34). This nationwide analysis reveals that PVD in ESKD patients is associated with increased hospitalization rates with cerebral ischemic events and reduced NIH events. Among KTx recipients, PVD correlated with increased hospitalizations for ischemic events, without affecting NIH. This highlights management concerns for patients with KTx and PVD.

Illicit kidney trade networks, operating globally, involve intricate interactions among various players, most notably buyers, sellers, brokers, and surgeons. A comprehensive understanding of these trade networks is, however, hindered by the lack of systematically amassed data for analysis. Further, extracting the geographic locations of buyers, sellers, brokers, transplant surgeons, and medical facilities in all relevant publications often involves extensive, time-consuming, manual labelling that is very costly. Although current techniques such as Named Entity Recognition (NER) tools can potentially automate the process, they are limited to identifying country names and often fail to associate the roles (i.e., offering buyer, seller, broker and/or surgery) that each country played. This study employed state-of-the-art technologies, including Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre-Trained Transformers (GPT) model Llama3.3 from Meta in developing a kidney trade country database. We first extracted news articles reporting illicit kidney trade from the LexisNexis database (2000–2022). BERT and Llama3.3 with chain-of-thought prompt tuning strategies were then applied to the materials to determine the relevance of articles to the illegal kidney trade and to identify the roles those different countries played in kidney trade cases over the past 23 years. The specific country classes recorded in the final kidney trade database included: a) countries of origin for kidney sellers; b) countries of origin of kidney buyers; c) countries performing illegal transplant surgeries; and d) countries of origin of organ trafficking brokers. The BERT classification model achieved an accuracy of 88.75%, ensuring that only relevant articles were analyzed. Additionally, the Llama3.3-70B model with chain-of-thought prompt tuning strategies extracted location-based roles with an accuracy of 86.30% for sellers, 88.89% for buyers, 93.33% for brokers, and 95.93% for surgeries, supporting these observed patterns. We observed in the final database that the kidney trade networks change and evolve dynamically where the primary role played by each country (as a host of either sellers, buyers or surgeries) change over time. About half of the top 10 countries playing each role gets replaced by other countries within a decade. The final database also demonstrated that developing countries were more likely to be a host of kidney sellers while that played by developed countries was a host of kidney buyers. The current study developed a geospatial database describing transnational kidney trade country networks over the past two decades. The new approach for geographic location extraction that is more precise compared to conventional NER and machine learning methods.