Unsupervised feature extraction of aerial images for clustering and understanding hazardous road segments.

Aerial image data are becoming more widely available, and analysis techniques based on supervised learning are advancing their use in a wide variety of remote sensing contexts. However, supervised learning requires training datasets which are not always available or easy to construct with aerial imagery. In this respect, unsupervised machine learning techniques present important advantages. This work presents a novel pipeline to demonstrate how available aerial imagery can be used to better the provision of services related to the built environment, using the case study of road traffic collisions (RTCs) across three cities in the UK. In this paper, we show how aerial imagery can be leveraged to extract latent features of the built environment from the purely visual representation of top-down images. With these latent image features in hand to represent the urban structure, this work then demonstrates how hazardous road segments can be clustered to provide a data-augmented aid for road safety experts to enhance their nuanced understanding of how and where different types of RTCs occur.

Treatment Pathways and Health Outcomes of German Patients with Chronic Graft-Versus-Host Disease After Allogeneic Hematopoietic Cell Transplantation: A Retrospective Health Claims Data Analysis.

BACKGROUND: Although chronic graft-versus-host-disease (cGvHD) is an important long-term complication after allogenic hematopoietic cell transplantation (allo-HCT) and is associated with increased healthcare resource utilization, real-world evidence is scarce. OBJECTIVES: The aim of the study was to evaluate survival of patients with cGvHD in Germany and to analyze hospitalization and treatment patterns. PATIENTS AND METHODS: Based on a German claims database with 4.9 million enrollees, a retrospective longitudinal analysis covering a 6-year period between 2013 and 2018 was conducted. Patients with allo-HCT in 2014 or 2015 (index date) and no record of transplantation or documentation of GvHD 365 days prior to index were included. Patients who subsequently developed a cGVHD were compared with those who did not develop a cGVHD within 3 years after index date. cGVHD cases were identified based on documented International Classification of Diseases, Tenth Revision (ICD-10) diagnosis and treatment algorithms. Since the onset of cGvHD is defined at 100 days after allo-HCT, only those alive beyond day 100 were considered in the survival analysis. Patients who did not survive the first 100 days after allo-HCT were censored to prevent a selection bias due to early mortality within patients without GvHD. Survival rates were plotted using the Kaplan-Meier estimator. The number of hospitalizations and average lengths of stay as well as treatment patterns were descriptively examined. RESULTS: Overall, 165 cGvHD patients were identified and compared with 43 patients without cGVHD. Short-term survival rates were better for patients with cGvHD; the 6-month survival probability was 95.8% for patients with cGVHD and 83.7% for patients without cGVHD. However, long-term survival was better in patients without GvHD; The 30-month survival probability was 65.5% for patients with cGVHD and 76.7% for patients without cGVHD. While overall 90% of cGvHD patients were hospitalized at least once, the share was only half for patients without GvHD (44%). 78.2% of patients with cGVHD received corticosteroids in combination with other predefined immunosuppressants. CONCLUSION: Findings from this study reveal a high disease burden associated with cGvHD. This underlines the high medical need for new interventional strategies to improve survival and morbidity after allo-HCT.

Persistence With Biologic Therapy and Associated Costs of Patients With Inflammatory Bowel Disease: A German Retrospective Claims Data Analysis.

Background: In recent years, biologic agents became a relevant and promising treatment option for inflammatory bowel diseases (IBDs). However, high treatment costs and moderate remission rates lead to a high interest in treatment persistence and corresponding economic consequences. Methods: A retrospective health claims data analysis was conducted including biologic naive patients diagnosed with IBD between 2013 and 2018. Observation points were at 12 and 18 months of follow-up, starting from the first biologic prescription. Nonpersistence was defined as either no further prescription or prescription of another biologic agent within the days of supply per original prescription. Biologic agents included were Adalimumab, Golimumab, Infliximab, Ustekinumab, and Vedolizumab. Results: In total, 1444 patients with IBD were included in this analysis, mostly treated with Adalimumab (46.9%) and Infliximab (39.9%) as their first biologic treatment. After 12 months, 72.2% of patients were still persistent with their initial biologic treatment with the highest shares for Infliximab (74%) and Vedolizumab (72.4%). 27.8% of patients were nonpersistent, mostly due to a switch of biologic agent (75.8%). Cox regression identified female, hospitalizations, and simultaneous prescriptions of corticosteroids and immunomodulators as risk factors for nonpersistence. Treatment costs per year were approximately 3000€ higher for nonpersistent patients (27,146€) than for persistent patients (23,839€), mostly due to inpatient treatment costs. Conclusions: The persistence of biologic therapy in this study was rather high at 72% after 12 months, while nonpersistence was mostly due to switches to other biologic agents. Lack of persistence is associated with increased cost, mostly due to nonbiologic medication and inpatient treatment.