We propose a data augmentation technique utilizing a Diffusion-based generative deep learning model to address the issue of data scarcity in skin disease diagnosis research. Specifically, we enhanced the Stable Diffusion model, a Latent Diffusion Model (LDM), to generate high-quality synthetic images. To mitigate detail loss in existing Diffusion models, we incorporated lesion area masks and improved the encoder and decoder structures of the LDM. Multi-level embeddings were applied using a CLIP encoder-based image encoder to capture detailed representations, ranging from textures to overall shapes. Additionally, we employed pre-trained segmentation and inpainting models to generate normal skin regions and used interpolation techniques to synthesize synthetic images with gradually varying visual characteristics, while having limitations for clinical use, this approach contributes to enhanced data diversity and can be used as reference material. To validate our method, we conducted classification experiments on seven skin diseases using datasets combining synthetic and real images. The results showed improvements in classification performance, demonstrating the effectiveness of the proposed technique in addressing medical data scarcity and enhancing diagnostic accuracy.

This study aimed to investigate the interface pressure (IP) of patients using a standard hospital mattress and polyurethane foam mattress as support surfaces and present cut‐off points for IP in patients who exhibited skin changes. A total of 189 inpatients enrolled from six general wards and three intensive care units at a Korean University Hospital. Skin changes were classified, and peak IP at the sacral and occipital regions was measured using a pressure scanner. Differences in IPs according to mattress type were analysed using independent t‐tests. The receiver operating characteristic curve was constructed to determine the cut‐off point, and the area under the curve with a 95% confidence interval was obtained using the Stata 15.1.program. The IP for a standard hospital mattress was significantly higher than that of a polyurethane foam mattress. The cut‐off points for IP at the sacral region were 52.90 and 30.15 mm Hg for a standard hospital mattress and polyurethane foam mattress, respectively. The cut‐off point for IP at the occipital region was 36.40 mm Hg for a polyurethane foam mattress. Using IP measurements to prevent pressure injuries is important and employ individualised interventions based on the cut‐off points for different support surfaces.

Background: As chronic wounds such as pressure injuries (PIs) are frequently colonized and can easily deteriorate into infection, it is important to reduce their bacterial load, for which antimicrobial dressings can be needed. This study aims to evaluate the efficacy of a 3% povidone-iodine (PVP-I) foam dressing compared to that of a silver foam dressing.Methods: This prospective non-inferiority study was conducted between 2016 and 2019 at three sites in South Korea. A total of 80 PI subjects were randomized to be dressed with either PVP-I foam (experimental group) or silver foam (control group) for up to 8 weeks.Results: Based on the Pressure Ulcer Scale for Healing (PUSH) tool, 25.0% of the experimental groups and 17.5% of the control groups (χ2=0.743, P=0.389) healed by more than 70%. The degree of reduction in wound size was analyzed using Image J, and the experimental and control groups decreased by 41.6%±35.3% and 49.7%±38.2% (t=–0.986, P=0.327), respectively. A Kaplan-Meier survival analysis to confirm the time to heal showed that if more than 30% of the PUSH score was healed, the time to heal was 27.0±9.3 days and 18.0±2.8 days in the two groups (χ2=3.225, P=0.073), respectively. The healing rates at 50 days were 85.8%±8.9% and 93.9%±5.7% in the two groups (P=0.073), respectively. There were no statistically significant differences between the groups in all results.Conclusion: This study demonstrated the non-inferiority of the 3% PVP-I foam dressing compared to the silver foam dressing for PI treatment.

Pressure injuries pose significant health risks, especially for the elderly, immobile individuals, and those with sensory impairments. These injuries can rapidly become chronic, making initial diagnosis important. Due to the difficulty of transporting patients from local health facilities to higher-level general hospitals for treatment, it is essential to utilize telemedicine tools, such as chatbots, to ensure rapid initial diagnosis. Recent advances in artificial intelligence have demonstrated potential for medical imaging and disease classification. Ongoing research in the field of dermatological diseases focuses on disease classification. However, the assessment accuracy of artificial intelligence is often limited by unequal class distributions and insufficient dataset quantities. In this study, we aim to enhance the accuracy of artificial intelligence models by generating synthetic datasets. Specifically, we focused on training models for Pressure Injury assessment using both real and synthetic datasets. We used PI data at a domestic medical university. As part of our supplementary research, we established a chatbot system to facilitate the assessment of pressure injuries. Using both constructed and synthetic data, we achieved a top-1 accuracy of 92.03%. The experimental results demonstrate that combining real and synthetic data significantly improves model accuracy. These findings suggest that synthetic datasets can be effectively utilized to address the limitations of small-scale datasets in medical applications. Future research should explore the use of diverse synthetic data generation methods and validate model performance on a variety of datasets to enhance the generalization and robustness of AI models for Pressure Injury assessment.

Background: With the ultimate goal of developing chatbot content to address consumer inquiries about pressure injuries (PIs), this study analyzed consumer perceptions of PI using big data.Methods: This study collected text data, with PI as the central word, from three search engines (Naver, Daum, Google) from January 2019 through December 2022, using Textom version 4.5. The words were refined through text mining, keyword analysis, and TF-IDF (term frequency-inverse document frequency) analysis. N-gram analysis and centrality visualization were conducted using Ucinet 6.0. The keywords and frequencies were clustered based on the frequency of words used in CONCOR (convergence of iteration correlation) analysis.Results: Consumers for PI showed a high perception of common sites for PI, concept of PI, healthcare facility for PI, PI products, PI care, PI-related life, and PI-related disease.Conclusion: Development of chatbot content customized to consumers’ needs, based on seven clusters associated with consumers’ perception of PI obtained through extensive data analysis with PI as the central word, is expected to make a significant contribution to improving consumers’ understanding of PI and enhancing the quality of PI management.