Angela Eunyoung Kwon

With the increase in insured patients and an aging population, managing the length of stay (LOS) for inpatients has become crucial for controlling medical costs. Analyzing the factors influencing LOS is necessary for effective management. Previous studies often used multiple or logistic regression analyses, which have limitations such as unmet assumptions and the inability to handle time-dependent variables. To address these issues, this study applied survival analysis to examine the factors affecting LOS using the National Health Insurance Service (NHIS) sample cohort data from 2016 to 2019, covering over 4 million records. We used Kaplan–Meier survival estimation to assess LOS probabilities based on sociodemographic, patient, health checkup, and institutional characteristics. Additionally, the Cox proportional hazards model controlled for confounding factors, providing more robust validation. Key findings include the influence of age, gender, type of insurance, and hospital type on LOS. For instance, older patients and medical aid recipients had longer LOS, while general hospitals showed shorter stays. This study is the first in Korea to use survival analysis with a large cohort database to identify LOS determinants. The results provide valuable insights for shaping healthcare policies aimed at optimizing inpatient care and managing hospital resources more efficiently.

Online reviews that consist of texts and images are an essential source of information for alleviating data sparsity in recommender system studies. Although texts and images provide different types of information, they can provide complementary or substitutive advantages. However, most studies are limited in introducing the complementary effect between texts and images in the recommender systems. Specifically, they have overlooked the informational value of images and proposed recommender systems solely based on textual representations. To address this research gap, this study proposes a novel recommender model that captures the dependence between texts and images. This study uses the RoBERTa and VGG-16 models to extract textual and visual information from online reviews and applies a co-attention mechanism to capture the complementarity between the two modalities. Extensive experiments were conducted using Amazon datasets, confirming the superiority of the proposed model. Our findings suggest that the complementarity of texts and images is crucial for enhancing recommendation accuracy and performance.

Background/Objectives: In Korea’s emergency medical system, when an emergency patient arises, patients receive on-site treatment and care during transport at the pre-hospital stage, followed by inpatient treatment upon hospitalization. From the perspective of emergency patient management, it is critical to identify the high death rate of patients with certain conditions in the emergency room. Therefore, it is necessary to compare and analyze the determinants of the death rate of patients admitted via the emergency room and generally hospitalized patients. In fact, previous studies investigating determinants of survival periods or length of stay (LOS) primarily used multiple or logistic regression analyses as their main research methodology. Although medical data often exhibit censored characteristics, which are crucial for analyzing survival periods, the aforementioned methods of analysis fail to accommodate these characteristics, presenting a significant limitation. Methods:Therefore, in this study, survival analyses were performed to investigate factors affecting the dying risk of general inpatients as well as patients admitted through the emergency room. For this purpose, this study collected and analyzed the sample cohort DB for a total of four years from 2016 to 2019 provided by the Korean National Health Insurance Services (NHIS). After data preprocessing, the survival probability was estimated according to sociodemographic, patient, health checkup records, and institutional features through the Kaplan–Meier survival estimation. Then, the Cox proportional hazards models were additionally utilized for further econometric validation. Results: As a result of the analysis, in terms of the ‘city’ feature among the sociodemographic characteristics, the small and medium-sized cities exert the most influence on the death rate of general inpatients, whereas the metropolitan cities exert the most influence on the death rate of inpatients admitted through the emergency room. In terms of institution characteristics, it was found that there is a difference in determinants affecting the death rate of the two groups of study, such as the number of doctors per 100 hospital beds, the number of nurses per 100 hospital beds, the number of hospital beds, the number of surgical beds, and the number of emergency beds. Conclusions: Based on the study results, it is expected that an efficient plan for distributing limited medical resources can be established based on inpatients’ LOS.