This paper presents an epidemiologic study of appendicitis in Taiwan over a twelve-year period. An analysis of the incidence in the low-income population (LIP) is included to explore the effects of lower socioeconomic status on appendicitis. We analyzed the epidemiological features of appendicitis in Taiwan using data from the National Health Insurance Research Database (NHIRD) from 2000 to 2011. All cases diagnosed as appendicitis were enrolled. The overall incidences of appendicitis, primary appendectomy, and perforated appendicitis were 107.76, 101.58, and 27.20 per 100,000 per year, respectively. The highest incidence of appendicitis was found in persons aged 15 to 29 years; males had higher rates of appendicitis than females at all ages except for 70 years and older. Appendicitis rates were 11.76 % higher in the summer than in the winter months. A multilevel analysis with hierarchical linear modeling (HLM) revealed that male patients, younger patients (aged ≤14 years), and elderly patients (aged ≥60 years) had a higher risk of perforated appendicitis; among adults, the incidence increased with age. Moreover, the risk of perforation was higher in patients with one or more comorbidities. LIP patients comprised 1.25 % of the total number of patients with appendicitis from 2000 to 2011. The overall incidence of appendicitis was 34.99 % higher in the LIP than in the normal population (NP), and the incidence of perforated appendicitis was 40.40 % higher in the LIP than in the NP. After multivariate adjustment, the adjusted hospital costs and length of hospital stay (LOS) for the LIP patients were higher than those for the NP patients. Appendicitis and appendectomy in Taiwan had similar overall incidences, seasonality patterns, and declining trends compared to numerous previous studies. Compared to NP patients, LIP patients had a higher risk of appendicitis, longer LOS and higher hospital costs as a result of appendectomy.

Background: With the rapid growth of healthcare services, health insurance fraud detection has become an important measure to ensure efficient use of public funds. Traditional fraud detection methods have tended to focus on the attributes of a single visit and have ignored the behavioural relationships of multiple visits by patients. Methods: We propose a health insurance fraud detection model based on a multilevel attention mechanism that we call MHAMFD. Specifically, we use an attributed heterogeneous information network (AHIN) to model different types of objects and their rich attributes and interactions in a healthcare scenario. MHAMFD selects appropriate neighbour nodes based on the behavioural relationships at different levels of a patient’s visit. We also designed a hierarchical attention mechanism to aggregate complex semantic information from the interweaving of different levels of behavioural relationships of patients. This increases the feature representation of objects and makes the model interpretable by identifying the main factors of fraud. Results: Experimental results using real datasets showed that MHAMFD detected health insurance fraud with better accuracy than existing methods. Conclusions: Experiment suggests that the behavioral relationships between patients’ multiple visits can also be of great help to detect health care fraud. Subsequent research fraud detection methods can also take into account the different behavioral relationships between patients.

Health insurance fraud is becoming more common and impacting the fairness and sustainability of the health insurance system. Traditional health insurance fraud detection primarily relies on recognizing established data patterns. However, with the ever-expanding and complex nature of health insurance data, it is difficult for these traditional methods to effectively capture evolving fraudulent activity and tactics and keep pace with the constant improvements and innovations of fraudsters. As a result, there is an urgent need for more accurate and flexible analytics to detect potential fraud. To address this, the Multi-channel Heterogeneous Graph Structured Learning-based health insurance fraud detection method (MHGSL) was proposed. MHGSL constructs a graph of health insurance data from various entities, such as patients, departments, and medicines, and employs graph structure learning to extract topological structure, features, and semantic information to construct multiple graphs that reflect the diversity and complexity of the data. We utilize deep learning methods such as heterogeneous graph neural networks and graph convolutional neural networks to combine multi-channel information transfer and feature fusion to detect anomalies in health insurance data. The results of extensive experiments on real health insurance data demonstrate that MHGSL achieves a high level of accuracy in detecting potential fraud, which is better than existing methods, and is able to quickly and accurately identify patients with fraudulent behaviors to avoid loss of health insurance funds. Experiments have shown that multi-channel heterogeneous graph structure learning in MHGSL can be very helpful for health insurance fraud detection. It provides a promising solution for detecting health insurance fraud and improving the fairness and sustainability of the health insurance system. Subsequent research on fraud detection methods can consider semantic information between patients and different types of entities.

Although numerous epidemiological studies on appendicitis have been conducted worldwide, only a few studies have paid attention to the effect of socioeconomic status on appendicitis, particularly studies focusing on the low-income population (LIP). We analyzed the epidemiological features of appendicitis in Taiwan using data from the National Health Insurance Research Database from 2003 to 2011. All cases diagnosed as appendicitis were enrolled. Between 2003 and 2011, 2,916 patients from the LIP and 209,206 patients from the normal population (NP) were diagnosed with appendicitis. Our finding revealed that the ratios of comorbidities, complicated appendicitis, and readmissions in LIP patients were slightly higher than those of NP patients. LIP patients were more likely to live in suburban and rural areas, and hence a higher proportion of them were hospitalized in a district or regional hospital compared with NP patients. The crucially finding was that the overall incidence ratios of appendicitis, acute appendicitis, and perforated appendicitis in the LIP were substantially higher than those in the NP (36.25%, 35.33%, and 37.28%, respectively). The mean LOS in LIP patients was longer than that of NP patients. The overall case-fatality ratio of appendectomy in the LIP was higher when compared with the NP (0.41% versus 0.12%, p < 0.05). We also observed that appendicitis was occurred frequently in male patients, with a higher incidence for those aged 15–29 years in both the LIP and NP. The incidences of incidental appendectomy showed a decreasing trend in both the LIP and NP. Finally, a valuable discovery was that the total hospital cost was comparable between the laparoscopic appendectomy (LA) and open appendectomy (OA) (1,178 ± 13 USD versus 1,191 ± 19 USD, p < 0.05) in LIP patients because they saved more hospitalization costs than NP patients when the previous one chose the LA. This study confirmed that a lower socioeconomic status has significantly negative impact on the occurrence and treatment of appendicitis and appendectomy. In terms of hospital costs and LOS, LIP patients benefit more from the LA approach than they do from the OA approach in the treatment of appendicitis.

Controversy surrounding the role of percutaneous cholecystostomy (PC) is fed by the absence of large amounts of data concerning its outcomes, and many authors have maintained that there is no evidence to support a recommendation for PC rather than cholecystectomy (CCS) in elderly or critically ill patients with acute cholecystitis (AC). We conducted this study by tracking trends in the utilization and outcomes of PC and CCS using longitudinal health research data in Taiwan. Analyses were conducted on 236,742 patients, 11,184 of whom had undergone PC and 225,558 of whom had undergone CCS. Average annual percentage changes (AAPCs) from 2003 to 2012 increased significantly by 18.34% each year for PC and by 2.82% each year for CCS. The subset analyzes showed that the mortality rates were far higher in patients underwent PC than in patients underwent CCS in all subgroups, which increased from a minimum of 1.45-fold to a maximum of 34.22-fold. The gap of the mortality rates between PC group and CCS group narrowed as the patients aged and with the seriousness of the diseases increased. Most patients with PC or CCS who died in-hospital or within 30 days after discharge were 70 years of age or older, and a large number of them received a CCI score of at least 1. The AAPCs of the overall mortality rates from 2003 to 2012 decreased by 6.78% each year for PC and by 7.33% each year for CCS. PC was related to a higher rate of cholecystitis recurrence and readmission for complications, but a lower rate of in-hospital complications and routine discharge than CCS, and 36.41% of all patients treated with PC underwent subsequent CCS. Additionally, the patients with PC experienced longer hospital stays and generated higher costs than the patients with CCS. Patients who underwent PC demonstrated poorer prognoses than did patients who underwent CCS. The role of PC in the Tokyo guidelines may be overstated; it is not as safe as the Tokyo guidelines have suggested in moderate-grade cholecystitis cases, and it should be limited to only the elderly and sicker patients.

Drug–drug interaction (DDI) prediction has received considerable attention from industry and academia. Most existing methods predict DDIs from drug attributes or relationships with neighbors, which does not guarantee that informative drug embeddings for prediction will be obtained. To address this limitation, we propose a multitype drug interaction prediction method based on the deep fusion of drug features and topological relationships, abbreviated DM-DDI. The proposed method adopts a deep fusion strategy to combine drug features and topologies to learn representative drug embeddings for DDI prediction. Specifically, a deep neural network model is first used on the drug feature matrix to extract feature information, while a graph convolutional network model is employed to capture structural information from the adjacency matrix. Then, we adopt delivery operations that allow the two models to exchange information between layers, as well as an attention mechanism for a weighted fusion of the two learned embeddings before the output layer. Finally, the unified drug embeddings for the downstream task are obtained. We conducted extensive experiments on real-world datasets, the experimental results demonstrated that DM-DDI achieved more accurate prediction results than state-of-the-art baselines. Furthermore, in two tasks that are more similar to real-world scenarios, DM-DDI outperformed other prediction methods for unknown drugs.